AR I ESt/5SMiTHSON]AN2lNSTITUTION"NOIlfUllSNrNVmOSHllWS“'s3 I a VH 3 I

HI1SNI NVINOSHilWS S3iavaan LIBRARIES SMITHSONIAN INSTITUTIO r~ z: r- _ z: *2 . _•

ol

' m ^ rn " m

co ' £ co £ co

ARIES SMITHSONIAN INSTITUTION NOUnilJLSNI NVINOSH1IIAIS S3IHVB

CO z * CO Z CO

^ > ' S x^ost^' >

HI JLSNI *" NVINOSHII WS°° S 3 I a VH a n^LI B R AR I ES^SMITHSONIAN INSTITUTIO

~ CO ~ CO ~

uj g

O Xfoos ~ O *"

AR I ES*~$MITHSON IAN "^INSTITUTION NOlinillSNI ^NVINOSHIMS S3 I d VH 9

vy.p.c^/ ^ rn <o m

£ co £ co \ £

nuiSNi nvinoshiii^s s3iHvaan libraries smithsonian_jnstitutic

nilXSNl ^NVINOSHillAIS S3ldVd8l*l LIBRARI E S^ SMITHSONIAN^ INSTITUTIC

OT

S Hi

70

y's.'m* k w^/

CO *v Uy.fa\' V pp|

(/> ” co £ co .

VARIES SMITHSONIAN INSTITUTION NOIlfUllSNI NVINCSH1IINS S3 1 a Va 9

JaV^v^

-i

o X ,/'SsTS O .«Sk^W X /Pi

07 X J—

f'9N > W 2 v\ > x^ua^ 2

I ES ^SMITHSON IAN INSTITUTION NOlinillSNrNVINOSHilWS^Sa I UVh B 07 X co ~ . co

UJ TZ W fr v UJ

O o

LSNI^NVINOSHJLIIAIS S3 I HVd a II LI B RAR I ES^SMITHSONIAN^INSTITUTION t r~ X r- z . IZ 2

O /55i5n\ O AgV$fas /^Q^oX O

P3

o\ 33 - l$E7\' r^\'-Z.\ —% ™ ' ^kS-\ “5

q I / t ? yjyyr* %s/'- >», It*®- <aa^| — ’

1 > ( £5® 3<i V- .x^A > -^§<J j—

73 UJ

— vv:

m " x^vas^ m n^vosjav>x ^ m

CO — CO _ CO

IES SMITHSONIAN INSTITUTION NOilfUIlSNi NVINOSHJLIWS S3 lavaa 11 l

CO Z V CO 2 CO Z

S < a 2 2 . <

2 ' \T' >' 2 X^os^7 >v 2

2 60 Z CO Z CO

LSNI_ NVINOSHillNS S3IHVHail LIBRARIES SMITHSONIAN INSTITUTION f z \ 5 „ f? _ 5 ^

O XV\osv^X O 00

IES SMITHSONIAN INSTITUTION NOlinillSNI^NVINOSHliiMS^ S3 I H Vd a 11 ~*t ^ r“ ^ ^ 2 r~ 2

<T^50/G>v rr ~-- — ^ ~~ ay/ ir

^ </> m "^2^ £ m

= t» £ — c/>

iSNi nvinoshiiws ssiavaan LIBRARIES SMITHSONIAN institution

Z CO Z ,♦>. (/) -jr CO

1 „v/.'l A S 408%. I A

2 X^LT-/ > "^^0^ 5 2

Z * Z ' rn

IES SMITHSONIAN INSTITUTION NOilfUIlSNi NVIN0SH1IWS SHIHVaSll

^ co _ . . CO 21

“ '* i§§&\ - ^ /ss^jx w

™ /^'A ^ tfe? k /fc/ ^ '

<

O X^V pc^ 2 5 °

LSNI NVINOSHlIl^S S3 I dVd 8 11 LI B RAR I ES^SMITHSONIAN^INSTITUTION _ I <“ x- z r- Z r- 2

C3 ✓^T^stit7/5\ *’*’**’ CD or JvqNl

o^ff^oX H

rn ' xjvAsn^ m x/^osv^ m

CO _ CO E CO —

1 ES SMITHSONIAN INSTITUTION NOlifUllSNI NVINOSHillNS SBiaVHan L S , . z « z w 5

PSYCHE

A Journal of Entomology

Volume 80 1973

Editorial Board

Frank M. Carpenter, Editor W. L. Brown, Jr.

E. O. Wilson B. K. Holldobler

P. J. Darlington, H. W. Levi J. M. Burns R. E. SlLBERGLIED

Published Quarterly by the Cambridge Entomological Club Editorial Office: Biological Laboratories 1 6 Divinity Avenue Cambridge, Massachusetts, U.S.A.

The numbers of Psyche issued during the past year were mailed on the following idates:

Vol. 79, no. 4, December, 1972: April 25, 1973 Vol. 80, no. 1-2, March-June, 1973: September 7, 1973 Vol. 80, no. 3, September, 1973: December 16, 1973

L J

Q. L

^rL\

y'fv r

PSYCHE

A JOURNAL OF ENTOMOLOGY

Vol. 80

March-June, 1973

Nos. 1-2

CONTENTS

Notes on the Life Cycle and Natural History of Parides areas mylotes (Papilionidae) in Costa Rican Premontane Wet Forest. A. M. Young 1

Body, Web-building and Feeding Characteristics of Males of the Spider Araneus diadematus (Araneae: Araneidae). R. Ramousse 23

Annotations on Two Species of Linyphiid Spiders Described by the Late Wilton Ivie. P. J. van Helsdingen 48

Correlation Between Segment Length and Spine Counts in Two Spider Species of Araneus (Araneae: Araneidae). L. D. Carmichael 62

Ant Larvae of Four Tribes: Second Supplement (Hymenoptera : For- micidae: Myrmicinae). G. C. Wheeler and J. Wheeler 70

A New Species of Anacis from Northwest Argentina (Hymenoptera, Ichneumonidae). C. C. Porter 83

Growth of the Orb Weaver, Araneus diadematus, and correlation with Web Measurements. J. Benforado and K. H. Kistler 90

The Cockroach Genus Calolampra of Australia with Descriptions of New Species (Blaberidae) . L. M. Roth and K. Princis 101

CAMBRIDGE ENTOMOLOGICAL CLUB

Officers for 1972-1973

President H. F. J. Nijhout, Harvard University

Vice-President T. F. H la vac, Harvard University

Secretary R. B. Swain, Harvard University

Treasurer F. M. Carpenter, Ilarvard University

Executive Committee A. F. Newton, Jr., Harvard University

J. W. Truman, Ilarvard University

EDITORIAL BOARD OF PSYCHE F. M. Carpenter (Editor), Fisher Professor of Natural History , Harvard University

P. J. Darlington, Jr., Professor vf Zoology, Emeritus , Harvard University

W. L. Brown, Jr., Professor of Entomology , Cornell University ;

Associate in Entomology , Museum of Comparative Zoology E. O. Wilson, Professor of Zoology , Harvard University H. W. Levi, Professor of Biology and Curator of Arachnology , Museum of Comparative Zoology

H. E. Evans, Alexander Agassiz Professor of Zoology, Harvard University

J. M. Burns, Associate Professor of Zoology, Elarvard University

PSYCHE is published quarterly by the Cambridge Entomological Club, the issues appearing in March, June, September and December. Subscription price, per year, payable in advance: $4.50 to Club members, $6.00 to all other subscribers. Single copies, $2.00.

Checks and remittances should be addressed to Treasurer, Cambridge Ento- mological Club, 16 Divinity Avenue, Cambridge, Mass. 02138.

Orders for missing numbers, notices of change of address, etc., should be sent to the Editorial Office of Psyche, 16 Divinity Ave., Cambridge, Mass. 02138 For previous volumes, see notice on inside back cover.

IMPORTANT NOTICE TO CONTRIBUTORS Manuscripts intended for publication should be addressed to Professor F. M. Carpenter, Biological Laboratories, Harvard University, Cambridge, Mass. 02138.

Authors are expected to bear part of the printing costs, at the rate of $13.50 per printed page. The actual cost of preparing cuts for all illustra- tions must be borne by contributors: the cost for full page plates from line drawings is ordinarily $12.00 each, and the full page half-tones, $20.00 each; smaller sizes in proportion.

The December, 1972, Psyche (Vol. 79, no. 4) was mailed April 25, 1973

The Lexington Press. Inc.. Lexington. Massachusetts

PSYCHE

Vol. 80 March-June 1973 No. 1-2

NOTES ON THE LIFE CYCLE AND NATURAL HISTORY OF

PA RIDES ARCAS MYLOTES (PAPILIONIDAE) IN COSTA RICAN PREMONTANE WET FOREST*

By Allen M. Young Department of Biology, Lawrence University Appleton, Wisconsin 5491 1

The “ Aristolochia- feeding” swallowtails of the New World tropics comprise a well-known group of butterflies famous for their roles in mimicry complexes (Brower and Brower, 1964). Although the adult stages of many congeneric species of notable genera such as Battus and Parides have been known for some time (Godman and Salvin, 1879-1901; Seitz, 1924), there is considerably less informa- tion concerning the immature stages of these butterflies. This is particularly the case for the Central American species of Parides , one of the three genera ( Battus , Parides , and the Old World Troides) of the Troidini, the tribe of pharmacophagous swallowtails (Ehrlich and Raven, 1965). While the Troidini are most abundant in the Old World tropics, it is apparent that New World genera in this tribe, such as Battus and Parides , have undergone extensive speciation in Central and South America. And with the exception of a few studies such as the recent study of Battus polydamus in Costa Rica (Young, 1971a) and another on the related Ornithoptera alexandrae on New Guinea (Straatman, 1971), the life cycles, behavior, and food plants of many species remain obscure. It is believed that the primarily neotropical distribution of the Aristolochiaceae (Pfeifer, 1966) is a major factor in accounting for the extensive adaptive radiation of Parides and Battus on these plants ( Brower and Brower, 1964; Ehrlich and Raven, 1965).

It is the close and perhaps coevolutionary association of genera such as Parides with Aristolochia (in the Aristolochiaceae) and the co-occurrence of several sympatric congeneric species in lowland

* Manuscript received by the editor March 26, 1973

I

2

Psyche

[March-June

tropical forests (Young, 1971b) that makes these butterflies suitable candidates for the study of butterfly community structure in the tropics. In the Caribbean premontane wet forests of Costa Rica, there occur at least three species of Parides whose adults are often found together on the same flowers in forests: P. areas mylotes, P. childrenae, and P. sadyattes. Another subspecies of P. areas , namely mycale , is also seen in association with these species. As an initial approach to determining the ecological factors responsible for the co-occurrence of these similar species as a functional Mullerian mimicry complex (Young, 1971b), studies have been conducted on the life cycle, food plants, and other aspects of butterfly biology, for all of these species in Costa Rica. To date, the biological data for P. areas mylotes (Bates) both in the laboratory (Young, 1972a) and held (Young, 1971b; 1972a) has been the most extensive for these species. This paper touches upon various aspects of biology in this species not covered in the previous studies. Other reports will subsequently appear concerning the biology of the remaining species. Godman and Salvin (1879-1901) mention that P. areas mylotes is common in the Pacific and Caribbean lowlands of Central America, ranging from southern Mexico to Costa Rica. Thus the widespread geographical distribution of the butterfly throughout Central America makes it an even more attractive species to study from the standpoint of the effects of local selection pressures on natural history and life cycle.

Methods

The studies summarized here are: habitat selection, life cycle, larval food plant acceptance, and behavior of immatures and adults. Life cycle and larval food plant acceptance were examined in the laboratory, while the other studies were conducted in the field at two localities. At various times between late 1968 and mid- 1970, field studies of P. areas mylotes were conducted at Finca la Selva, a region of relatively undisturbed premontane tropical wet forest (elev. about 90 m) located on the confluence of the Rio Puerto Viejo and Rio Sarapiqui. During the months of July and August 1972, the butter- fly was studied at Finca Tirimbina, a forest site located about 8 km west of Finca la Selva and at the basal belt transition zone (about 200 m. elev.) between montane and premontane tropical wet forest.

Habitat selection was studied by observing feeding and egg-laying activities of adults at various places in the forest, both at “La Selva” and “Tirimbina”. At La Selva, habitat selection was studied spo-

1973]

Young — Parides areas mylotes

3

radically several days each month over a 14-month period. At Tirim- bina, it was studied systematically 14 days over a two-month period.

Life cycle studies consisted of the description of life stages and the estimation of egg-adult developmental time under “laboratory” conditions. These measurements were made on individuals reared on a natural food plant, and eggs were obtained in one of two basic ways. The first method was to collect eggs witnessed to be ovi- posited in the wild ; this method was employed primarily in the Tirimbina studies and to a lesser extent in the earlier La Selva studies. The second method was to obtain eggs by hand-pairing newly-emerged adults, using the technique of Clarke (1952) for Papilio machaon, or allowing mating to occur in pairs of adults confined to plastic bags. The latter technique is useful to obtain estimates of fecundity in this species (Young, 1972a). Both methods, obtaining eggs in the wild, and mating females in the laboratory with subsequent induction of oviposition, are very successful for this species, provide large numbers of eggs for rearing studies. Combining both methods, a large number of individuals were reared from La Selva (primarily through the laboratory mating method) and a lesser number were reared from Tirimbina. The “laboratory” for the La Selva studies consisted of a well-ventilated room in an apartment in San Jose, while the “laboratory” for the Tirimbina studies was a room in a different apartment, located about 1.5 km from the first. In both cases, air temperature usually varied between 2i-23°C and the humidity was about 45%.

The techniques for rearing immatures of this butterfly are given in Young (1972a) for La Selva individuals, and essentially the same methods were employed for the Tirimbina studies.

The larval food plant acceptance studies were conducted from individuals obtained at Tirimbina during 1972. This study consisted of offering first instar larvae immediately after hatching, in the laboratory, fresh clippings of several species of Aristolochia from various sources. The rationale was to offer separate small groups of young larvae various species of Aristolochia including species known to be natural food plants. Larvae on each food plant were then scored for survival rate and body size. There were five species of Aristolochia that were called “novel” food plants in addition to the two natural food plant species. Two experiments were con- ducted in San Jose: in each of these, 12 larvae were reared on the natural food plant and 13 were reared on each of two “novel” food plants collected from different localities in Costa Rica. The remain- ing three food plants were tested at Lawrence University during

4

Psyche

[March-June

September and October 1972. The three species of Aristolochia involved were already growing in a greenhouse tropical room for about two years, and the Parides eggs were transported (by air) from Costa Rica to Lawrence on September 6, 1972. Since the natural food plant was not in culture at Lawrence, enough cuttings of it were also brought to Wisconsin to sustain the larvae through the earlier instars. At Lawrence, 10 larvae were on the natural food plant, and 8 on each of the “novel” food plants.

Field studies of larval and adult behavior consisted of making repeated observations on the feeding, resting, and defensive habits of larvae in different instars, and on the oviposition behavior of adults.

Results

1 Habitat selection

Adults of both sexes of P. areas mylotes are most commonly en- countered along paths, natural clearings, swamp edges, and other exposed areas that either border forest or those which are found in the forest interior. For example, the general study site at Tirimbina where adults were most frequently seen is between the edge of forest and a small river (Fig. 1). This small strip of dense secondary growth vegetation is the result of forest being cut back from the river edge for the original purpose of growing yucca and other veg- etables that form the major diet of these people. Here, the adults fly low over dense second-growth vegetation, seldom crossing the small river, and frequently flying several meters into the shaded forest understory and canopy. Excursions into the forest were most frequently done by mated females in search of oviposition sites while males and very fresh (presumably unmated) females generally lin- gered in the sunlight second-growth. The strip of second-growth between the forest and river is a major courtship area for this butter- fly at Tirimbina and extensive growths of the larval food plant vines are found hanging down from trees along the forest edge and grow- ing horizontally in the canopy within a few meters from the edge. A later paper (Young, et ah, in prep.) will demonstrate that mated females of this species are far more prone to dispersal than either males or unmated females. In the present paper, we can say that mated females cruise along extensive tracts of cleared forest edge in search of egg-laying sites, while males and unmated females remain close to their eclosion sites. Courtship encounters are generally con- fined to low sunlight vegetation very close to where the adults emerged from their pupae. Males precede females in emergence.

1973]

Young — Parides areas mylotes

5

Fig. 1. A major habitat of adult Parides areas mylotes (Bates) at Finca Tirimbina, near La Virgen, Heredia Province, Costa Rica. An adult popu- lation is found along the forest edge, and males are active in the low secondary growth vegetation between the forest and small river (Rio Tirimbina) to the left. August, 1972.

Thus habitat selection, which can obviously be exercised only by the adults (since eggs and larvae are relatively fixed through the oviposi- tion strategy), is molded strongly in this species by two factors: ( i ) establishment of optimal courtship sites by males in sunlight second-growth bordering forests or forest clearings, and (2) the response by mated females to become more prone to disperse in search for oviposition sites. Similar adult movement patterns have been seen at La Selva, and the lesser vagility of individual males was mentioned in Young (1971b). A courtship strategy in which males patrol an area of the habitat consistently day after day (Young, et al., in prep.) and mate with females as the latter emerge from their pupae, is optimal for butterflies in which males are shorter-lived than females, as is the case with Parides (Young, 1972a). But Cook et al., (1971) report a short life expectancy of about 10 days in P. anchises

6

Psyche

[March-June

and P. neophilus in a seasonal habitat on Trinidad where torrent rains may kill off the adults of both sexes.

Although adult feeding preferences do not appear to be ,a major factor in dispersal patterns at Tirimbina, it is interesting to note that mimetic association with other species of Parides is most intense at nectaries at La Selva (Young, 1971b). At Tirimbina, P. areas mylotes is the only species of this genus seen consistently at the study site, and flower specificity is not apparent. At La Selva., this butterfly as a functional component of Mullerian mimicry complexes exercise a strong preference to visit a single species of flower {I~I amelia patens ) also visited by other Parides (Young, 1971b) ; judging from the amount of time spent daily at Hamelia flowers, there appear to be very few or no other preferred adult food sources of Parides at La Selva. In the absence of the other Parides at the Tirimbina study site, adult P. areas mylotes is found on a variety of flowers, usually ranging from red to purple. Thus in the absence of strong selection pressures favoring mimetic association, and where this mimicry is potentially most effective at a food source, flower specificity may break down for Parides in habitats where the species do not co-occur on a regular basis. Similar diurnal patterns of visitation at flowers between members of a tropical Battus mimicry complex in addition to the co-occurrence of several Parides at flowers at La Selva suggest strong selection pressures resulting in convergence of feeding habits to enhance mimicry (Young, 1971b; 1972b).

Although courtship activity is generally limited to the sunniest hours of the morning (Young, et ah, prep.), adults of both sexes and various age-classes (distinguished by the extent of wing tatter- ing) generally forage throughout the day, and they are relatively unaffected by changes in local weather conditions. Even at a mon- tane tropical forest locality (Cuesta Angel) where a cloud forest occurs at about IOOO meters elevation, adults are seen foraging throughout the day at the bright red flowers of Impatiens sultani ( Balsaminaceae — “Touch-me-nots”), a small herbaceous plant that is imported from Africa and that grows in large numbers. As the day becomes less bright in terms of illumination from the sun, these flowers become even more conspicuous due to increased contrast of the red coloration with the misty air; to the human observer, the flowers are more conspicuous, and perhaps the butterflies respond in a similar fashion. In both lowland and mountain localities, adult activity drops off sharply after about 4:00 P.M. When there is short succession of unusually dry days in both lowland and mountain

1973]

Young — Parides areas mylotes

7

localities, adults, especially males, are frequently seen visiting reced- ing mud puddles and moist patches of ground.

Life cycle and developmental time

The egg (Fig. 2-A-C) is deep rusty-brown and slightly flattened at the base. The diameter is i.i mm and the egg is covered with an irregular thick layer of an orange-red sticky substance, which at- taches it to the leaf surface and gives the entire surface of the egg a rough appearance. This sticky substance forms thin threads which hang down from the upper half of the egg and assist in attachment (Fig, 2-A). It is not known if the sticky substance is also defensive in function, in the sense of discouraging attack by ants and other leaf-wandering predatory arthropods. The apical region of the egg darkens considerably immediately before hatching. Eggs are gen- erally laid on the ventral surface of older leaves and occasionally in the crotches of small stems and petioles (Fig. 2-C). The amount or thickness of the sticky substance covering the eggs is apparently very variable, since other details of egg external morphology, such as deep grooves (Fig. 2-B, C) can be seen on some eggs while com- pletely obscured on others. Eggs are laid singly but usually in loose clusters of 2-5 eggs on a single leaf.

At La Selva, the natural food plant is tf Aristolochia sp.” (this is a new species from northeastern Costa Rica soon to be described by H. W. Pfeifer based on my collection of it during March, 1970). At Tirimbina, the natural food plant is Aristolochia constricta Griseb. Both of these species occur in lowland forest on the Carib- bean side of the central Cordillera in Costa Rica. Pfeifer (1966) mentions that A. constricta is a forest species found from Costa Rica to Panama, the Lesser Antilles, and probably northern South America.

The first instar is about 3.2 mm long when it hatches, and the ground color of the body is dark orange-brown. The head is shiny black. After the young larva begins to feed on leaf tissue, the body ground color becomes a deep wine red. All segments bear long tubercles of the same color as the body, but the lateral pair on the first segment are orange-white, and this color also characterizes the dorsal pairs of tubercles on segments two, seven, ten, and twelve (Fig. 2-D ) . The tubercles are fleshy for about one-third their length, with the apical two-thirds being stiff and bearing numerous tiny black spines (Fig. 2-D). The oSmeterium is bright orange- yellow throughout larval life. By the time of the first molt, the larva is about 9 mm long.

8

Psyche

[March-June

I ! ■ 1 jflj

IlllMMiiiMil

1973]

Young — Parides areas mylotes

9

The second instar is remarkably similar in appearance to the first instar, with the only major difference being a loss of the spines seen on tubercles in the previous instar. The larva (Fig. 2-E, F) retains the six rows of spines of the first instar, in addition to the shiny black head and true legs. The precise arrangement of the tubercles is very noticeable in this instar. The first four thoracic segments bear two pairs of lateral tubercles, and the uppermost pair disappears until the tenth segment where it is resumed until the twelfth seg- ment. The two pairs of lateral tubercles on these segments are not precisely in line: the tubercle of thoracic segment i are juxtaposed with those of thoracic segment 2 etc. The lateral tubercles of the thoracic segment i are considerably shorter than these tubercles on the remaining segments. The dorsal pair of tubercles on abdominal segments i and 4 are white, while the upper lateral pair of the fourth and fifth abdominal segments are also white. The highly reduced dorsal pair of the abdominal segment 1 1 are also white. This pat- tern of tubercle arrangement and coloration is retained throughout the rest of larval life. By the second molt, the larva is about 14 mm long.

The third instar is an exact replica of the second instar except that the ground color of the body is a very deep purplish black. The third instar is shown in Fig. 2-G. By the time of the third molt, the larva is about 23 mm long. The fourth instar (Fig. 3-A) is identical to the third instar except that the skin is very shiny and reflective. It attains a length of 35 mm by the fourth molt.

A dramatic change in the ground color occurs with the molt to the fifth instar (Fig. 3-B, C). The ground color is a dull, velvety purplish-brown mottled with irregular blotches of black. The black coloration is most extensive on the segments bearing white tubercles (Fig. 3-B, C). The wfrite ridge along the anterior edge of the osmeterial cuff behind the head is more prominent in this instar. As this instar continues to grow, the ground color becomes even lighter in coloration as extensive velvety grayish-tan areas replace the for- merly purplish-brown areas of the body. The coloration of the dark tubercles is also variegated during the fifth instar, with each tubercle bearing lines of white in addition to the mottled coloration of the

Fig. 2. Life cycle and behavior of Parides areas mylotes (Bates). (A) dorsal view of two eggs on a leaf; note the rough surface and sticky strands on the eggs (B) single egg showing deep vertical grooves (C) sin- gle egg in crotch of stems (D) first instar, lateral view (E) two second instar larvae (one is feeding) (F) several second instar larvae living to- gether (G) third instar, dorsal view.

IO

Psyche

[March-June

1973]

Young — Parides areas mylotes

1 1

body ground color. By the time of pupation, the larva is about 45 mm long. The coloration of the larva remains unchanged at the time of pupation.

The pupa (Fig. 3-D) is about 25 mm long and the color pattern consists of various light shades of green and yellow. The frontal portions of the thorax and abdomen are yellow while the rest of the body is light green.

Godman and Salvin (1879-1901) and Seitz (1924) give good il- lustrations of wing color patterns of the adults (Fig. 3-E). The single light area of the dorsal surface of the forewing in the female is cream-colored while the dorsal bands on the hindwings are orange- red. This color pattern is very consistent in both laboratory-reared and wild-caught females of P. areas mylotes. Less stable is the fore- wing dorsal coloration in the male within a single local population. The large spot on each forewing (Fig. 3-E) is light green but with the apical portion being cream-colored. Considerable variation is apparent in this “two-component’ 5 spot on the dorsal surface of the male’s forewing; this variability concerns the presence, absence, and size of a second, very small two-component spot just inside the radial cell of each forewing, and almost touching the major spot (Fig.

3-E). Similarly, there is considerable variation in the discal cell

spot. Godman and Salvin (1879-1901) mention the considerable variation in the forewing spotting pattern of male in the closely related species, P. iphidamas. Adults of both sexes of P. areas mylotes can be distinguished from the subspecies mycale by the presence of a thin light red marginal border of the wings in the former subspecies, while these markings are white in the latter sub- species. The red patch on the dorsal surface of the hindwings in

male P. areas mylotes is more intense than in the female, and the distribution of the coloration is very different between the sexes (Fig. 3-E). In bright sunlight, the red patches of the male’s hind- wing are often iridescent, giving off a purple lustre; this is not seen in the female. The mean length of the forewing in the female is about 40 mm, while the same statistic of the male is about 38 mm. Thus, not only is there a striking color sexual dimorphism in this butterfly, but also a consistent wing length difference between the sexes. In the absence of crowding, laboratory-reared individuals often

Fig. 3. Life cycle and behavior of Parides areas mylotes (Bates). (A) fourth instar, lateral view (B) fifth instar, lateral view (C) fifth instar, feeding on the tip of a young stem of Aristolochia (D) pupa, lateral view (E) adults, female above, male below.

12

Psyche

[March-June

bear the same wing-length as wild-caught individuals from the same locality.

The egg-adult developmental time for P. areas mylotes in the laboratory for individuals reared on Aristolochia constricta is sum- marized in Table i. In a previous study (Young, 1972a), the egg- adult developmental time of this butterfly on Aristolochia sp. from La Selva was about 42 days. The developmental time in that study was measured on eggs obtained from La Selva adults. The develop- mental time for eggs obtained at Tirimbina, and reared on A. con- stricta is 53 days (Table 1). This difference in developmental time between the two populations is apparent in eggs, larvae, and pupae: the egg stage lasts 4 days in La Selva individuals as opposed to 6 days in Tirimbina individuals ; the total larval period for La Selva in- dividuals is 17 days as opposed to 33 days in Tirimbina individuals; the pupal stage lasts 21 days in La Selva individuals as compared to 14 days in Tirimbina individuals.

Larval food plant acceptance

Development from the egg stage on natural food plants is suc- cessfully completed in the laboratory (Young, 1972b; Table 1). When other species of Aristolochia are tested, differences in food plant acceptance by the larvae become apparent. Development is successfully completed, and without a change from the Tirimbina developmental time when larvae are reared from the egg stage on Aristolochia labiata Willd. in Costa Rica. But larvae die during the first instar when offered A. veraguensis Duchr. in Costa Rica. For

Table 1. The developmental time of Parides areas mylotes on a natural food plant, Aristolochia constricta, under laboratory conditions.*

		INSTAR	INSTAR	INSTAR	INSTAR	IN STAR	TOTAL
	EGG	1	2	3	4	5	PUPA EGG-ADULT
MEAN DURATION (days)	6	5	5	6	6	11	14 53
± S.E.	± 0.1	± 0.3	± 0.5	± 0.3	± 0.2	± 0.8	± 0.2
N	46	46	42	42	40	37	37

^Laboratory conditions consisted of confining larvae to closed plastic bags containing clippings of food plant. Physical conditions around the bags were 21-23 °C and about 45% relative humidity. See text for further details of rearing techniques, laboratory conditions, etc.

1973]

Young — Parides areas mylotes

13

the rearing studies at Lawrence, all the larvae died either in the first or second instar when reared on A. ringens Vahl, A. littoralis Parodi, and A. gigantea (Mart. & Zucc.). For the groups of larvae offered these species, survivorship was 0%. Thus, in addition to the two known natural food plants of P. areas mylotes , namely Aristo- lochia sp. from La Selva and A. constricta from Tirimbina, the butterfly only feeds successfully on A. labiata Willd. in Costa Rica.

Behavior of adults and larvae

Observations on adult behavior are limited to the oviposition strat- egy of this species, since a later report (Young et al., in prep.) will discuss other aspects of adult behavior, most notably, the spacing patterns of males and females, and the courtship strategy.

Adults of both sexes generally cruise very low over second-growth vegetation. Mated females in search of oviposition sites exhibit extreme forms of cruising behavior in two ways : ( 1 ) they perform sudden, almost vertical darts into the canopy where Aristolochia lianas are found, and (2) they flutter through very dense second growth within a few inches of the ground, and often being obscured from view for several minutes.

Such patterns of cruising behavior by egg-laying females are con- sistent with the observation of well-developed food plant specializa- tion in this butterfly. The usual situation locally is that eggs are laid on a single species of Aristolochia , and there is considerable site- selectivity exercised in terms of placing the individual eggs securely on the older leaves of an individual plant. The eggs are seldom laid on young leaves and occasionally on stems at crotches between two stems. Eggs are customarily laid on the dorsal surface of older, well-shaded leaves of the vine, and anywhere from one to five eggs may be laid in a loose cluster in this manner. Upon landing on a leaf for oviposition, the female exhibits considerable wing fluttering and drumming behavior with the antennae; an egg is usually laid within 12 seconds. Oviposition is most commonly seen during sunny hours throughout the day. While males may be cruising in the general vicinity of egg-laying females, there is virtually no observable interactions between the sexes. The less cohesive nature of the mated female portion of a local breeding population of P. areas mylotes (Young et al., in prep.) results in there usually being no more than one or two ovipositing females at a larval food plant patch on a given day. These individuals cover large tracts of habitat in searching for oviposition sites, but usually return repeatedly on the same day to a given food plant patch.

While clustering of eggs in the field is generally loose and vari-

14

Psyche

[March-June

able, when mated females of this butterfly lay eggs in the laboratory, there is usually a tight clustering of eggs (Young, 1972a). Thus tight clustering of eggs (the arrangement of eggs into a group where the eggs touch each other) can be induced in the laboratory when females are confined individually or in low numbers to plastic bags containing clippings of the food plant. Such clustering, however, is seldom found in the wild in this butterfly and other species of Parides.

The larvae of P. areas mylotes exhibit several behavioral patterns that warrant more intensive study. Upon hatching the larva in- variably eats its emptied egg shell, and then moves a considerable distance to the closest youngest leaves. Locomotor movement is ac- companied by the production of silken treadwork on which the larva crawls from one place to another. Although small groups of larvae are frequently found in the field (Fig. 2-E, F) there is no evidence for gregarious habits among the individuals in a group. All individ- uals on an individual vine generally feed at the same times of day, but there is no indication of coordinated locomotor movements among the individuals. Furthermore, single or doublets of larvae are also frequently encountered in the field. Larvae of all instars are gen- erally inactive at night. The extent of larval dispersion when several eggs are laid on a vine may be governed by the size of the vine. For example, it is not uncommon to find one or two fourth or fifth instar larvae present on a young vine ( 1-2 m tall) in the field, and in cases where there are two present, these individuals are often found to- gether on the same stem. Both in the field and laboratory, older larvae eat the stems of young Aristolochia vines (Fig. 3-C). On very large vines in which woody tissue is well-developed, older larvae are generally confined to feeding on leaves and it is unusual to find two or more individuals resting close together. Group formation is frequently encountered only in the younger larvae (first and second instars) and in cases where larger (older) larvae are clumped, this is most likely due to the fact that they are feeding on a young vine and the food supply is limited. It is not known if Parides larvae crowded on young Aristolochia vines will leave the vine in response to intense crowding. The osmeteria of the larvae of swallowtail butterflies are functional defense organs. Predatory attack on the larvae of P. areas mylotes in the wild has not been observed to date. The defensive strategy of the larvae against predators includes ( 1 ) possession of conspicuous body coloration in which the dark body and pattern of white tubercles stands out against the light green coloration of Aristolochia leaves, (2) possession of an apparently

1973]

Young — Parides areas mylotes

15

functional and brightly-colored defensive organ, the osmeterium, and (3) probably the possession of generally toxic or poisonous systemic properties making the insect unpalatable, since they feed on vines reputed to have very toxic properties.

Discussion

Young (1971a) reported a developmental time for Battus poly- damus on Aristolochia veraguensis of about 14 days under similar laboratory conditions to those employed in the present study. Straat- man (1971) reported the developmental time of Ornithoptera alex- andrae Rothschild to be 13 1 days on Aristolochia schlechteri and 107 days on A. tagala, where the difference occurred during the larval period. The developmental time of Parides areas mylotes on Aristolochia sp. from La Selva is 42 days (Young, 1972a) while 53 days on A. constricta (Table 1.). Furthermore, the develop- mental time of Parides childrenae on Aristolochia pilosa at La Selva is about 42 days (Young, 1972a). Thus different genera in the Troidini have different developmental times on different species of Aristolochia. At La Selva there has been ecological divergence be- tween P. areas mylotes and P. childrenae with respect to the species of Aristolochia used for oviposition and larval food-consumption. Furthermore, two different strains of P. areas mylotes are evolving between La Selva and Tirimbina: the duration of all immature life cycle stages has been altered and the species feeds on a different species of Aristolochia at each locality. If this difference in develop- mental time was due solely to differences between the larval food plant species, we would expect to find only a change in duration of the larval period similar to that noted by Straatman (1971) in Ornithoptera alexandrae on New Guinea. But in the case of P. areas mylotes, there has been a change in the embryonic and post- embryonic developmental time which suggests genetic alterations. Strain-effect is not solely confined to food plant differences of the type noted for Victorina epaphus on the Pacific and Caribbean slopes of the central Cordillera in Costa Rica (Young, 1972c). Precisely what sorts of ecological factors are reshaping the developmental architecture of P. areas mylotes at different localities on the Carib- bean drainage of the central Cordillera in Costa Rica remain obscure at this time. One interesting hypothesis concerning this question would focus on a higher level of predation pressure on eggs and larvae in La Selva populations of the butterfly, which would favor an accelerated developmental period for these life stages.

i6

Psyche

[March-June

The inability of young larvae of P. areas mylotes to survive on Aristolochia ringens, A. littoralis, A. gigantea, and A. veraguensis may be due to the lack of evolutionary contact (Ehrlich and Raven, 1965) with these plants. An alternative explanation is that extreme food plant specialization in the butterfly has resulted in the narrow restriction to only a few species of Aristolochia locally. Until more is known about the regional and geographical distribution of various species of Aristolochia in Central America, it will be difficult to resolve the question of larval food plant adaptability in Parides. Unfortunately eggs from La Selva have not been reared on A. con- stricta from Tirimbina nor the converse, namely, eggs from Tirim- bina reared on A ristolochia sp. from La Selva.

The question of unpalatability is of considerable ecological and evolutionary interest. Brower and Brower (1964) have demon- strated that freeze-killed adult Parides neophilus L., which feeds on various species of Aristolochia on Trinidad, are very unpalatable to Scrub Blue Jays in the laboratory. Brower and Brower (1964), Ehrlich and Raven (1965) and Pfeifer (1966) cite previous studies which illustrate the toxic properties of various compounds derived from the vegetative portions of Aristolochiaceae. The question of palatabifity in genera of the Troidini ( Parides , Battus , Ornithop- tera, and Troides ) is of interest since the larvae are presumably unpalatable in addition to possessing a defensive organ (Eisner et. al., 1971). The larvae of these genera, as exemplified in the present study by P. areas mylotes , are generally conspicuous in appearance (Fig. 2, 3) to the human observer.

The possession of a dual system of defense by Parides larvae and other troidines is related to the functional responses of each com- ponent (unpalatability and chemical defense secretion) to different kinds of predators that the larvae encounter in their habitats. Un~ palatability, as evidenced here by the conspicuous coloration of the larvae and the toxic properties of their food plants, is an adaptation for defense against vertebrate predators such as insectivorous birds, mammals, and reptiles. Brower and Brower (1964) have demon- strated that blue jays become ill after eating an unpalatable butterfly and that there is a subsequent modification in prey-selection behavior by such an experienced predator to avoid the prey on further visual contact with it. Thus, the flexible learning abilities of vertebrate predators makes unpalatability an effective defensive mechanism that increases the likelihood of survival of individuals in a prey popula- tion. An insectivorous bird foraging in forest edge second-growth or forest canopy has daily opportunity for visual contact with the

1973]

Young — Parides areas mylotes

7

poisonous Parides larvae which stand out against the foliage back- ground during the day time when they are feeding. This is an ideal situation for unpalatability to be effective against vertebrate preda- tors. The bird does not have to make tactile contact with the poten- tial prey, but can recognize it from a distance. On the other hand, the added possession of a defensive organ that produces a volatile chemical secretion would be an adaptation primarily against inverte- brate (arthropodan) predators that make tactile or very close visual contact with Parides larvae and elicit a behavioral response. Such a defensive mechanism would be essentially ineffective against verte- brate predators since the larvae could not respond fast enough to the strike of the predator, and the larva would invariably be killed. This is especially true since lepidopterous larvae have low visual sensing ability but quick discriminatory ability for tactile stimuli. In a similar fashion, the generally instinctive nature of the be- havioral repertoire of invertebrate predators would make unpalata- bility an ineffective defense mechanism against these predators. Under such conditions, there is strong selection for the evolution with a dual system of defense, one adapted to vertebrate predators with developed learning abilities (unpalatability), and the other adapted to smaller invertebrate predators with instinctive behavior patterns (defense glands). Furthermore, the larvae would probably survive single attacks by invertebrates such as ants, even though the in- stinctive nature of the predator’s behavior results in repeated attacks on the prey. The small size of invertebrate predators and the ability of Parides larvae to survive individual attacks (in the form of small bites) reduces the threat of death from instinctive predatory be- havior patterns. Thus, in the absence of conclusive evidence, I sug- gest that the unpalatable properties of troidine butterfly larvae (Euw et al., 1968) are an adaptation to potential large vertebrate preda- tors, while their defensive organs comprise an adaptation to inverte- brate predators. This effect is even more pronounced in the adults, which are very unpalatable to birds (Brower and Brower, 1964), since there are ample opportunities for foraging birds which catch insects on the wing to recognize, at a distance, the butterflies through conspicuous coloration. Therefore, adult butterflies should possess unpalatability rather than defensive gland as an adaptive strategy against vertebrate predators. The studies of Euw et al. (1968) and Eisner et al. (1971) indicate that unpalatability and chemical de- fense secretions in troidine butterflies are due to very different kinds of chemical compounds.

The oviposition behavior varies greatly for different genera of

i8

Psyche

[March-June

troidine butterflies. Straatman (1971) found that Ornithoptera alex- andrae lays eggs singly, and Cook et al. (1971) comment that single oviposition also occurs in Parides neophilus and P. anchises on Trini- dad. But Young (1971a) found tight cluster oviposition in the held to prevail in Battus poly damns in Costa Rica. The oviposition in P. areas mylotes is very variable since eggs may be laid singly or as loose clusters of varying numbers of eggs per cluster. But oviposi- tion in the wild is never tightly clustered as seen in Battus polydamus (Young, 1971a). The P. areas mylotes pattern is basically single, but with a behavioral tendency to lay several eggs close together on a single leaf. This behavior results in first and second instar re- maining together in small groups and dispersing later, which is very different from the more well-defined gregarious behavior exhibited by the larvae of Battus polydamus (Young, 1971a). Larvae in the latter case are generally gregarious through all instars and presumably fitness is increased as noted in other studies (Ghent, i960). A similar oviposition pattern to that found in P. areas mylotes also occurs in P. childrenae and P. sadyattes (Young, in prep.). Thus the oviposition pattern of Parides in Costa Rica (and perhaps for all of the Central American mainland) is a variable one being basically single but typified by loose clusters of a variable number of eggs, usually ranging between two and five on a leaf. It is clearly not entirely single, nor is it the tight-cluster- ing pattern seen in Battus. As might be predicted, the larvae are semi-gregarious in P. areas mylotes (Fig. 2-E, F) as well as in P. childrenae and P. sadyattes (Young, in prep.) and perhaps in most Parides , while truly gregarious in Battus. These preliminary find- ings in different species suggest that there may exist distinct phylo- genetic patterns of type of oviposition and extent of larval gregari- ousness at the generic level in the Troidini, and perhaps within other tribes of Papilioninae. Superimposed upon evolutionary history will be the prevailing ecological conditions (Birch and Ehrlich, 1967) such as food plant specialization, patchiness of food plant populations, predation pressure on immatures, adult population cohesiveness, and several others, which mold the oviposition strategy in either direction (single versus clustering) and the likelihood of larval gregarious behavior.

Summary

In this paper concerning the life cycle and natural history of Parides areas mylotes (Bates) on the Caribbean side of the central Cordillera in Costa Rica, the following points were emphasized:

1973]

Young — Parides areas mylotes

19

( 1 ) The butterfly is a forest species which is most commonly encountered along forest edges associated with extensive borders of secondary growth vegetation or small forest clearings.

(2) Habitat selection by adults is governed primarily by two factors: (a) the selection of optimal courtship sites by males ex- hibiting home range behavior, and (b) the search pattern of mated females for suitable oviposition on Aristolochia vines along forest borders and in the canopy.

(33 The larvae of this species are probably warningly-colored, since they contrast greatly with the light green leaves of the food plant. The pupae are cryptically colored against the same back- ground.

(4) The egg-adult developmental time varies on different natural food plants in different localities: on Aristolochia sp. from Finca La Selva the developmental time is about 42 days; on A . constricta from Finca Tirimbina 53 days. This difference is due to more than food plant difference since the egg stage is considerably shorter in individuals reared on Aristolochia sp. There appears to have been the evolution of different strains in different localities where different food plants are also exploited.

(5) Development is successfully completed on A. labiata but unsuccessful on A. veraguensis , A. ringens , A. littoralis, and A. gigantea. The inability of young larvae to feed on these species may be due to either (a) a lack of contact with those species, or (b) the development of narrow food plant specialization.

(6) The conspicuous coloration (contrast) of the larvae against the light green food plant leaves and the known toxic properties of the Aristolochiaceae indicate that the larvae are unpalatable to verte- brate predators with well developed learning abilities. The un- palatability of the larvae is inferred from the known unpalatability of the adults of a related species of Parides. The possession of an osmeterial defensive organ is interpreted here, on the other hand, as being primarily an adaptation of defense against invertebrate (ar- thropodan) predators with rather inflexible (instinctive) learning abilities.

(7) The variable oviposition strategy of P. areas mylotes in the wild is not strictly single nor is it clustering. Eggs are generally laid in loose clusters of two to five eggs on a leaf, and this pattern appears to be a modified form of single oviposition. When mated females are confined to plastic bags in the laboratory, tight clustering of eggs can be induced. Previous studies show that at least one tropical species of Battus lays eggs in tight clusters in the wild,

20

Psyche

[March-June

while some species of Parides undoubtedly lay eggs singly and Ornithoptera lays eggs singly. It is suggested that there may exist phylogenetic differences in oviposition patterns at the generic level in the Troidini, and that secondary differences in these patterns are molded by contemporary ecological factors.

Acknowledgements

The La Selva portion of these studies was financed by N.S.F. Grant GB-7805, Daniel H. Janzen, principal investigator, with logistic support through the Organization for Tropical Studies, Inc. The Tirimbina studies were financed by Grant No. 12 1 of the Bache Fund of the National Academy of Science and partially by N.S.F. Grant GB-33060. Logistic support was provided by the Costa Rican program of the Associated Colleges of the Midwest. Roger Kimber and John Thomason assisted with rearing and food plant acceptance studies. Howard W. Pfeifer identified the species of Aristolochia and provided rooted cuttings of several species. Lee D. Miller confirmed the identification of the butterfly. The manu- script was read by Murray S. Blum.

Literature Cited

Birch, L. C., and P. R. Ehrlich.

1967. Evolutionary history and population biology. Nature 214: 349- 352.

Brower, L. P. and J. V. Z. Brower.

1964. Birds, butterflies, and plant poisons: a study in ecological chem- istry. Zoologica 49: 137-159.

Clarke, C. A.

1952. Hand pairing of Papilio machaon in February. Entomol. Record 64: 98-100.

Cook, L. M., K. Frank, and L. P. Brower.

1971. Experiments on the demography of tropical butterflies. I. Sur- vival rate and density in two species of Parides. Biotropica 3 : 17-20.

Ehrlich, P. R., and P. H. Raven.

1965. Butterflies and plants: a study in coevolution. Evolution 18: 586-608.

Eisner, T.

1970. Chemical defense against predation in arthropods. In Chemical Ecology (ed. by Sondheimer, E. and Simeone, J. B.), pp. 157-218. New York: Academic Press.

Eisner, T., A. F. Kluge, M. I. Ikeda, Y. C. Meinwald, and J. Meinwald.

1971. Sesquiterpenes in the osmeterial secretion of a papilionid butter- fly, Battus polydamas. J. Insect Physiol. 17: 245-250.

1973]

Young — Parides areas mylotes

21

Euw, J. V., T. Reichstein, and M. Rothschild.

1968. Aristolochic acid-1 in the swallowtail butterfly Pachlioptera aristolochiae (Fabr.) (Papilionidae) . Israel J. Chem. 6: 659-670. Ghent, A. W.

1960. A study of the group-feeding behavior of larvae of the Jack pine Sawfly, N eodiprion pratti banksianae Roh. Behavior 16: 110-148. Godman, F. D. and O. Salvin.

1879-1901. Biologia centrali-americana, Insecta, Lepldoptera-Rhopalo- cera. Vol. I.

Jordan, K.

1907-1908. Papilio. In The Macrolepidoptera of the world. Vol. 5. The American Rhopalocera. (Ed. by A. Seitz), pp. 11-45. Stuttgart: A. Kernan Verlag.

Pfeifer, H. W.

1966. Revision of the North and Central American hexandrous species of Aristolochia (Aristolochiaceae) . Annals Missouri Botan. Gar- den 53 : 115-196.

Straatman, R.

1971. The life history of Ornithoptera alexandrae Rothschild. J. Lepid. Soc. 25: 58-64.

Young, A. M.

1971a. Mimetic associations in natural populations of tropical papilionid butterflies. I. Life history and structure of a tropical dry forest breeding population of Battus polydamus polydamus. Rev. Biol. Trop. 19: 211-240.

1971b. Mimetic associations in natural populations of tropical papilionid butterflies (Lepidoptera : Papilionidae). J. New York Entomol. Soc. 79 210-224.

1972a. Breeding success and survivorship in some tropical butterflies. Oikos 23: 318-326.

1972b. Mimetic associations in populations of tropical butterflies. II. Mimetic interactions of Battus polydamus and Battus bellus. Biotropica 4: 17-27.

1972c. The ecology and ethology of the tropical nymphaline butterfly, Victorina epaphus . I. Life cycle and natural history. J. Lepid. Soc. 26: 155-170.

BODY, WEB-BUILDING AND FEEDING CHARACTERISTICS OF MALES OF THE SPIDER ARANEUS DIADEM ATUS (ARANEAE: ARANEIDAE)

By Raymond Ramousse*

Division of Research

North Carolina Department of Mental Health P. O. Box 7532 Raleigh, North Carolina 27611

INTRODUCTION

Many investigators have observed female orb-web spiders in their natural habitats (Enders, 1972; Eberhard, 1971), but there have been relatively few scientific observations of males outdoors. A major reason for this is because after maturation males discontinue web- building and they seek mates and are difficult to follow in an un- confined setting. Males have also attracted less attention in labora- tory situations since they have shorter life spans than females and because they stop building webs after reaching maturity. The activity of spiders in laboratories has been observed primarily in relation to their web-building behavior (LeGuelte, 1966, Witt, 1963a, b), making the female a more frequent subject of study. Thus, with the exception of maturation on web-building (Witt et al.f 1972), only females have been comprehensively studied.

The focus of this research is to explore the activities of the males of Araneus diadematus ’Clerck and their role in the female-male relationship which ultimately determines the continuity of the species. Two characteristics related to the females have already been identi- fied as possibly playing a part in the survival of the species. These include cocoon hatching and differential maturing. Cocoons have been observed hatching at two different times for a single species of spider — presumably providing an advantageous distribution of egg- production over a period of time (Potzsch, 1963). Also, within a set1 of spiderlings, different rates of maturation have been observed. Some females grow rapidly and die early while others grow slowly

^Present address of author: Laboratoire d’Ethologie experimentale, 1 rue Raulin, 69 Lyon 7e, France

Manuscript received by the editor March 26, 1973

To avoid confusion with the designation of “family” used in nomencla- ture, offsprings from a single cocoo-n will be called a “set.”

22

1973]

Ra?mousse — A raneus diadematus

23

and live at least four months longer (Reed & Witt, 1972). The related differential maturing rates may provide an advantageous distribution of spiderlings over a period of time. Together, these mechanisms would seem to help a species survive drastic or poten- tially destructive changes in environmental conditions. This research seeks to explore the male’s role in these phenomena. At what rate is he growing, maturing and dying during the female’s life cycle?

This leads to the question of inbreeding. An observation of the maturation rates of spiderlings of the same set was conducted in an effort to determine if inbreeding is possible.

Also, if the rate of growth is a factor in the rate of maturation (and spiders of the same set are known to present a considerable variation in size even under apparently optimal conditions), (Witt et al., 1968), is growth prenatally or genetically determined or a function of external factors?

The effects of an even diet independent of manifest behavior (Witt et al. , 1972) and differential force-feeding on various schedules (Benforado & Kistler, 1972) have already been studied. What, however, would happen to the growth rate of male and female spiders if they could choose their food quantity through web-building frequency?

The answers to some of these questions about the growth and maturation of male spiders should provide clues about their role in the reproductive cycle and, more generally, about their role in the continuity of the species.

METHODS

Two Araneus diadematus cocoons collected in the field, were placed in two different rearing boxes in the laboratory, where they hatched (February 23, 1972, one cocoon and 14 days later, March 6, 1972, the other). The offspring from the first cocoon will be called set I, and the offspring from the second cocoon, set II. The laboratory provided a cycle of long warm days and short cool nights throughout the lifespan of the animals.

As the animals left the communal web to build individual webs, they were put in glass tubes. Five weeks after hatching for set I and three weeks after hatching for set II the spiderlings were caged in individual labeled frames (50 X 50 X 10 cm) where they could build webs without apparent limitation in size. All observations began at this moment; however, some molts were noticed inside the cocoon, and the spiderlings molted one or two times in the glass tubes.

24

Psyche

[March-June

In the rearing boxes as well as in the glass tubes they were pro- vided water and gnats ad libitum. In the frames the spiderlings were fed with de-winged houseflies. A weighed fly was given one time every three days only when a web had been built, thus rewarding the spiders for high frequency of building.

The individual weights of the spiders (accuracy o.i mg) were recorded every week and web-building was recorded every day. Each web was photographed then collapsed by the experimenter, and ana- lyzed for size, shape, fine structure and regularity (Reed et al., 1965). The dates of the molts of each spiderling were recorded and the length of the first leg was measured on the molted limb (ac- curacy in mm.).

In the following pages the initials FG and SG are used in place of fast growing males and slow growing males. Statistical com- parison between the two groups (SG & FG) where not specifically mentioned was made with the Wilcoxon test, adapted by White for unpaired measurements (White, 1952).

RESULTS

Of 31 spiderlings that reached maturity in set I, twelve were identified as males. There were 14 males out of a total of 29 animals in set II. The number of males in each set is significantly repre- sentative of the expected 50% probability of males in a population (Binomial test, p = 0.01 in each case).

Some characteristics of the male

The adult males of Araneus diadematus have enlarged black palps, relatively narrow elongated abdomens, and weigh about a fifth of the adult females. Adult females are characterized by long yellow palps and a globulous abdomen (Figure 1). Other characteristics of the males include banding of the legs that is generally darker, a lack of humps on the abdomen, and a modified second tibia that is stronger than in females and has short spines (Levi, 1971).

The enlarged palps appear at the end of the next-to-the-last molt, whitish instead of black, and blacken between the two last molts. One animal exhibited enlarged palps prematurely two molts before the last one and four other animals after the last molt, but these were exceptions.

After the last molt, when they reached sexual maturity and maxi- mum weight, the males stopped building webs. Sekiguchi (i955) reported that a male of Araneus ventricosus, in the laboratory, did

1973]

Ramousse — A raneus diadematus

25

Figure 1. Outline of a male (left) and a female (right). Note the difference of size (female front leg: 16 mm, male front leg: 12 mm), of weight (female: 144.1 mg, male: 47.0 mg) and the difference of form of the palps (short and enlarged for the male, long and thin for the female).

not spin a web after its last molt, and that the aggregate glands become vestigial in the adult males. Prior to this point the involve- ment of the aggregate glands in the formation of the catching area of a web was clearly shown (Peakall, 1964). We may suppose that adult males are unable to spin webs because their aggregate glands are no longer functional.

The males ate scarcely, even when we attempted to induce prey catching by placing the flies in front of their mouths. While an immature male transformed a fly into a small compact ball through eating; the different parts of the body of a fly abandoned after eating by a mature male were easily recognizable. Even when they ate, the mature males used only a small amount of the food available. Males of Linyphia triangularis Clerck did not require food in the adult stage, and were still able to mate with females that later produced fertile eggs. When these males were provided with food, the rate of prey capture and the rate of food consumption dropped sharply

26

Psyche

[March-June

Figure 2. Body weight of four FG and seven SG littermate males in set I of Araneus diadematus, hatched in the laboratory from one cocoon on February 23, 1972. Dashed line: weekly mean body weights of the FG males. Dotted line: weekly mean body weights of the SG males. Numerals followed by an arrow indicate the number of animals molting for the last time during a week. Numerals surmounted on black circles indicate the number of animals dying during a week. The FG males reached their maximum weight the 13th week of post-hatching, the SG males reached their maximum the 29th week of post-hatching. Note that the SG animals need twice as much time to mature as the FG.

(Turnbull, 1962). We may assume that the adult males, no longer able to build a web, do not neeed food to fulfill their mating role.

Four males in this study continued to spin webs until they died; they built webs for a few days after the last molt was recorded, then stopped building for three or four weeks and generally built a final web six or seven days before death. These facts suggest that these four males were not able to go through an additional molt to com- plete their development. Also, these males presented enlarged palps only after the last molt recorded which is another confirmation of thir inability to complete their development.

During the web building period the males are distinct from the females only between the two last molts (about 3 weeks). This explains why few studies have been made of the males either outdoors or in the laboratory.

1973]

Ramousse — A raneus diadematus

27

W eight increase

In each set, the individual weight curves follow two distinct pat- terns and no in between: a group with an early maximum (FG) and a group with a late maximum (SG). In set I the course of the growth of four males with early maxima (between 10th and 15th week post hatching) was compared to seven males with late maxima (between 22nd and 33rd week post hatching) (Figure 2). In the second set the growth of 1 1 males which reached their maximum weight between the 8th and 16th week post hatching was compared to three males reaching their maximum weight between the 19th and 23rd week post hatching (Figure 4). In both sets the SG animals needed approximately twice as much time to complete the last molt and to attain sexual maturity as did the FG animals. In each set the females could be divided into fast and slow growth groups in the same way as males. Figure 6 shows the body weight of the FG and SG males and females. The data from the two sets were combined forming four groups: FG and SG males and females. The weight gain per day until maturation, in both sets, was sig- nificantly higher for the FG males than for the SG males (set I: T = 6,P :: - 0.05; set II: T = 7, P = 0.05).

The mean weight gain per day between the two last molts for each group was :

set I set II

FG	2.09 mg/d	1.59 mg/d
SG	0.64 mg/d	1. 6 1 mg/d

In each set, every animal showed a weight gain per day significantly higher between the two last molts than during the preceding period of observation (Wilcoxon matched-pairs signed ranks test: set I: N = 9, T = 3, P = 0.02; set II : N feft 13, T = o, P — 0.01).

Frequency of building

Th® mean of webs built per day to reach the last molt were:

set I set II

FG	0.57 web/day	0.49 web/day
SG	0.22 web/day	0.18 web/day

week percentage web building

28

Psyche

[March-June

The FG males had a higher rate of building while they grew than did the SG males (set I: T = 6, P = 0.05; set II: T = 6, P — 0.01 ) . The differences in the rate of building appear clearly on the graphs (Figs. 3 and 5) obtained by plotting the mean fre- quency of building per week for each group in each set.

The frequency of building is strongly correlated with the amount of food eaten per day (Kendall rank coefficient; set I: 7 — 0.59, P = 0.004; set II: 7 — 0.52, P = 0.005). This is the necessary consequence of the feeding schedule. We might suppose that this relation occurs in nature. A fresh snare probably increases the chances of capturing prey.

F G males
'f?j set 1
F G males [ l 1
- 1 1 1 1 1 1 1 1 f 1 1 1 1 I". 1 1 1 1 r •. ; ; ; .
rj— J -J ;		.... j
— 1 • :	••••••	.....
• ...... . .r"*
! .

8 12 16 20 24

weeks post hatching

Figure 3. Frequency of building of the FG and SG males of set I. Dashed line: weekly mean of frequency of building for the four FG males. Dotted line: weekly mean of frequency of building for the seven SG males. Note the similarity in the pattern between weight increase and web building frequency. (Compare with Fig. 2.)

1973]

Ranvousse — A raneus diadernatus

29

Figure 4. Body weight of 14 male littermates in set II of Araneus diadernatus hatched in the laboratory on March 6, 1972. Dashed line: weekly mean body weight for the 11 FG males. Dotted line: weekly mean body weight for the three SG males. Numerals followed by an arrow indicate the number of animals molting for the last time during a week. The FG animals reached their maximum weight the 12th week post-hatch- ing, the SG males reached their maximum the 27th week post-hatching, when the FG males are dead. (Compare with Fig. 2.)

The rate of building :

set 1 set II

FG	0.57 w/d	0.64 w/d
SG	O.31 w/d	0.30 w/d

between the two last molts was significantly higher than the rate of building during the previous stages of growth in both sets (set I: N = 9, T = 2, P = 0.01 ; set II: N — 12, T = 1.5, P = 0.01 Wilcoxon test).

What explanations are there for differences in frequency of build- ing? A multiplicity of factors have been found to have some in- fluences on web-building: a change from dark to light, a steep rise

30

Psyche

[March-June

Figure 5. Frequency of building of the FG and SG males of set II. Dashed line: weekly mean frequency of building for the 11 FG males. Dotted line: weekly mean frequency of building for the three SG males. Note similarity to Fig. 3.

in temperature following a temperature minimum, weather condi- tions, barometric pressure, a full silk supply, hunger (Witt, et al ., 1968). In the laboratory, all the spiders were subjected to the same environmental conditions, therefore the differences in rate of building should be due to an internal state, such as hunger. There is a gen- eral agreement in the literature that hunger is a strong drive for web-building. Heavy feeding is followed by several days without web-building (Koenig, 1951; Wolf & Hempel, 1951; Wiehle, 1927; Peters, 1932). The interpretation is that the hunger drive is too low for releasers like temperature and light to operate. On the other hand, spiders deprived of food built almost every day (Peters, 1939) and built webs even at the expense of other body constituents (Witt, 1963b). We may assume that the FG males have a higher level of hunger than the SG males, which induces a higher rate of building.

Pood consumption

Each time a spider was fed, the fly was weighed before eating. Since only one or two percent of a fly was rejected by a spider after

1973]

Ra?nousse — A raneus diadematus

3i

eating, we assume that a fly was eaten entirely. The mean quantity of food consumed per day was :

set I set II

FG	2.44 mg/d	2.06 mg/d
SG	i.44mg/d	1.39 mg/d

The FG spiders ate a significantly higher quantity of food per day than the SG ones (set I: T = 6, P ~ 0.05; set II : T = 8, P — O.05 ) . There was a significant difference in the amount of food consumed per day between FG males of the two sets (T = 8.5, P = 0.05).

The mean quantity of food eaten between the last two molts was:

set I set II

FG	3-33 mg/d	2.61 mg/d
SG	2.85 mg/d	3.54 mg/d

In each set the mean quantity of food consumed per day between the last two molts was significantly higher than the mean amount of food eaten per day during the preceding observation period, (Wil- coxon test: set I : N = 10, T = o, P = 0.0 1 ; set II: N = 13, T = 1, P = 0.01).

A relationship exists between the amount of food eaten per day and the growth rate in both sets, indicating that the growth rate is a function of the amount of food consumed (Kendall rank coefficient; set I: y = 0.55, P = 0.01 ; set II : y = 0.60, P = O.OOi). The foot eaten was used to sustain the basal metabolism, to make silk, and to build the body of the spiders. A rough estimate of the per- centage of food transformed into spider tissues was obtained by dividing the gain of body-weight per day by the quantity of food consumed per day: the FG males used about 57 % (set I) and 47% (set II) of the food they ate, while the SG males transformed only 33% (set I) or 32% (set II) of their food into spider tissues. The FG groups transformed a greater amount of food consumed into spider tissues than did the SG groups (set I : T = 6, P == 0.05;

32

Psyche

[March-J une

Figure 6. Body weight and number of molts of 25 males (15 FG, 10 SG), and 25 females (15 FG, 10 SG) from the two sets cocoons of Araneus diadematus studied. Each line connects mean body weights at one, two, five, seven and nine months post-hatching. Large black circles: FG females, large dashed line: early life of SG females; small black circles: FG males, small dashed line: SG males. Arrows indicate the number of molts to the time. Note the different growth rates and the related different speed of maturation in FG and SG males and females, and the similarities for both sets.

set II: T = 14, P = O.05). As a result of having more food available for metabolism, an FG male was able to utilize more energy for other metabolic processes than basal metabolism, such as synthesis of silk, synthesis of body constituents, etc. This would assure a larger supply of silk for the FG spiders than for the SG, which could be an important drive for web-building (Peakall, 1967). The increased frequency of building in the FG spiders leads to a greater amount of food consumed which in time results in the rapid weight gain and growth.

1973]

Ranvousse - — A ranens diadematus

33

Maturation

Between the start of the observations and the time of sexual ma- turity (last molt) the mean number of molts recorded for each group was:

set I set II

FG	3.25 molts	3.27 molts
SG	4.50 molts	3.66 molts

The SG males in set I went through a significantly higher number of molts than did the FG males (T = 12, P = 0.05) and reached a higher weight (see below). In set II, we had only three SG males and one of them did not complete its development, this explains the difficulty to obtain a significant difference between SG and FG ani- mals in this set.

For set I the mean time of maturation was 81.6 days for the FG spiders and 202.5 days for the SG spiders. In set II maturation was reached in a mean time of 78.0 days for the FG males and 163.0 days for the SG ones. The time of maturation was significantly longer for the SG animals (set I: T = 6, P — 0.05 ; set II: T = 6, P = 0.01). In addition the time of maturation was sig- nificantly longer for the SG in the first set than in the second set (T = 6, P = 0.05) .

The rate of maturation, number of molts divided by the number of days necessary to complete these transformations, was significantly higher for the FG males than for the SG males (set I: T — 6, P = 0.0 1 ; set II: T — 6, P = 0.05).

The Kendall rank coefficient between the gain of weight per day and the number of molts per day was 0.61 for set I and 0.66 for set II ( in both P — 0.001). A relationship exists between the rate of growth and the rate of maturation which is in agreement with the findings of Deevey (1949) with Latrodectus mactans (Fabri- cius) and of Benforado and Kistler (1973) with Araneus diadema- tus. We may assume that the maturation rate is correlated with the growth rate. The mean length of time in days between two con- secutive molts was determined. In 3 out of 4 groups, the last inter- molt was longer than the other intermolts (table 1); for the FG males, this last intermolt was significantly longer than the earlier (N — 10, T = 1.5, P = 0.01 ).

34

Psyche Table i

[March-June

	1	2	3 4
set I FG	20.6	14.0
SG	63-3	37-5	46.1 29.6
set II FG	22.7	12.0
SG	26.5	54-0	55.0 17.0
Mean length of time	in days	separating two consecutive molts. The

numerals designate each intermolt and its order in relation to the final one, i being the last.

Increase in leg-length

The mean length of the first leg as measured on the last molt was :

set I set II

FG	1 1 .0 mm	10.5 mm
SG	14.3 mm	12.3 mm

SG males, which were also heavier, had significantly longer first legs than FG males after the last molt (set I: T — io, P = o.oi ; set II : T = 7.5, P = 0.05 ) .

The rate of leg growtdi is given by the ratio of the length gain in the number of days necessary to obtain this increase of length. The mean rate of leg growth during the entire observation was:

set I set II

FG	0.183 rnm/d	0.176 mm/d
SG	0.069 rnm/d	0.069 mm/d

The rate of length increase was significantly higher for the FG males than for the SG males (set I: T — 10, P = 0.01; set II: T = 7, P = 0.05). This points out the relationship existing be- tween rate of maturation and rate of lengthening. No correlation was found between the leg-growth between molts and the length of time of the intermolt.

Maximum weight

Body weight increased for all males to a maximum at the last molt, declining from this point onwards. This is in contrast to

1973]

Ramousse — A raneus diadematus

35

female weight increase which continues after the last molt, possibly due to egg formation. The mean weight was :

set I set II

FG	62.95 mg	57.64 mg
SG	88.52 mg	70.23 mg

The maximum weight reached by the FG males, in both sets, was lower than for the SG males. This suggests that the maximum weight may be a function of the duration of development. In that case, rapid maturation would occur at the expense of weight growth.

No correlation between initial weight and final weight was found in contrast to the findings of Benforado and Kistler (1972). A relatively small difference in initial weights and the low accuracy of the weights may explain the contrast.

Mating

The FG males of set I matured 81.6 days post-hatching (p-h.) and those of set II 78.0 days p-h. The FG females reached ma- turity 229 days p-h. in set I and 104 days p-h. in set II. In the first set all the FG males were dead before any of the FG females were mature, preventing them from mating. In the second set, the FG males survived 71 days after the last molt, so that some of them were still alive when the first FG females matured. But the females accepted the advances of the males only 60 days or more after the last molt, preventing the FG males from mating with their sisters. In summary, the FG males of both sets were sexually mature too early to mate with any of the females of the same set.

The SG males reached maturity 202 days p-h. in set I and 163 days p-h. in set II. At this time the FG females of set II were already mature (104 days p-h.) and those of set I were almost ma- ture (229 days p-h.), as well as the SG females of set II (223 p-h.). In these conditions the SG males of both sets may have been able to mate with the FG females of their own set or the other set.

Each of the nine SG males, when they were an average of 300 days p-h., were brought into the presence of three different females. All the males seemed to behave in the same way, but only three of them mated successfully with a single female, and one with two different females. These successful males were the biggest of the SG males. One male of each set was able to mate with a female of

3 6

Psyche

[March-June

his own set. Only the FG females of the set I accepted the males, while both FG and SG females of set II accepted the males. How- ever, the small number of males limited the number of trials and did not permit us to know statistically which of the females (FG or SG) were the most successful in mating.

The FG males cannot mate with females of their own set, but we may assume that they can find females of other sets in a natural habitat which are mature at the same time. The SG males can mate with the FG females of their own set, permitting limited in- breeding. This is merely a possibility since different sets in nature may have mature females at the same time.

Comparison of the webs of SG and FG spiders

FFeb changes during development :

Webs for the FG and SG males were compared on the basis of the spiral area, mesh size and thread length.

The spiral area of all webs built by males in both groups and both sets showed a general increase until reaching a maximum area during the period between the last two molts. The spiral area decreased in size thereafter. Four males mentioned earlier, who did not follow the general web-building pattern (they built webs until they died), showed no decrease in spiral area in their final webs. This further supports speculation that they died before achieving full development through a complete series of molts. In contrast, the spiral area of the females of Araneus diadematus and Neoscona vertebrata in- creased until the last three months, after which time the catching area does not change significantly (Witt & Baum, i960). The catching area of the females of the golden garden spider, Argiope aurantia , showed a growth and decline, the peak size coinciding roughly to the time of last molt and sexual maturation (Reed, Witt & Scarboro, 1969).

An essentially upward linear growth in mesh size throughout the lifetime occurred for 11 of the males studied. For the 12 other males the mesh size increased until reaching a plateau during the last intermolt. Witt, Rawlings and Reed (1962) have pointed out that the mesh size of the female webs of Araneus diade?natus show also an increase until the last molt, and then reach a plateau. But Argiope aurantia shows a linear growth in mesh size throughout the lifetime (Reed, Witt & Scarboro, 1969).

The thread length fol’ows the same pattern as the two other pa- rameters with a peak during the last intermolt and then a decrease. Argiope aurantia and Araneus diadematus females have been shown

1973]

Ranvousse — A raneus diadematus

37

Figure 7. Web built by a young male of Araneus diadematus. Young FG and SG males built webs with the same characteristics, small and fine- meshed. The vertical white lines of the scale are spaced 20 mm apart.

38

Psyche

[March-June

to change according to the same pattern for thread length (Reed, Witt & Scarboro, 1969; Witt, Rawlings & Reed, 1972). The de- crease of the thread length may be due to the thickening of the threads as the weight of the spider increases (Christiansen et al., 1962). The males follow the same pattern of web-changes as the females of the same species, except for the catching area. The gen- eral effect is that young Araneus diademcitus males build small, fine-meshed webs (Fig. 7) ; and during the onset of the last inter- molt build large, wide-meshed webs; males at the end of the last intermolt, without changing weight and leg-length significantly, build medium meshed webs. Therefore it appears that web size cannot simply be explained by the spider’s bodily dimensions (Witt, Rawl- ings & Reed, 1972; Reed, Witt & Scarboro, 1969).

Comparison of the webs of the FG and SG males at the same age

All the webs photographed between the 9th and 12th weeks post hatching of five FG males of both sets were measured. So were the webs of eight SG males of both sets during the same period of time. All these spiders were at the same age, but the five FG spiders were almost mature and the eight SG males were 100 days before reach- ing maturation. Table 2 shows the figures for spiral area, mesh size, thread length and a measure of the regularity of the spacing of the threads (standard error of median mesh size North). The two measures of the body, and the four measures of the web show sig- nificant differences between the FG and SG males, with the excep- tion of the variance of the mesh size: the webs of the two groups show similar regularity.

Table 2

	FG	SG	t	P
Body weight	50.1 ± 19.0 mg	22.6 ± 6.4 mg	3.58	0.005
Leg length	9.1 ± 1.70 mm	6.68 ± 1.48 m	2.49	0.05
Spiral area	42,138 ± 9,883 mm2	13,956 ± 8,178 mm2	4.88	0.001
Mesh size	58.60 ± 6.96 mm2	34.75 ± 10.08 mm2	4.05	0.005
Thread length	16,066 ± 4,111 mm	6,950 ± 3,282 mm	4.07	0.005
SE median mesh size North	0.114 ± 0.030	0.162 ± 0.062	1.47

Measures of webs of the FG and SG males at the same age. Only the regularity measures in the last line are not significantly different.

1973]	Ranvousse — A raneus diadematus		39
	Table	3
	FG	SG	t	P
Body weight	60.14 ± 11.56 mg	79.37 ± 18.07 mg	0.11
Leg size	10.75 ± 1.22 mm	13.60 ± 1.41 m	4.90	0.001
Spiral area	33,931 ± 10,131mm2	28,766 ± 6,559 mm2	1.15	0.400
Thread length	15,546 ± 2,768 mm	10,425 ± 1,715 mm	3.66	0.005
Mesh size	46.46 ± 9.64 mm2	70.73 ± 15.95 mm2	3.99	0.005
SE median mesh size	0.900 ± 0.022	0.293 ± 0.068	8.78	0.001

Measure of webs of the FG and SG males at comparable stage of maturity.

It is neither possible to relate the regularity measures to leg length — the FG’s legs were significantly longer than the SG’s legs — nor to maturation since the FG spiders were at the last stages of matura- tion and the SG males two or three stages before. We may assume that the size of the males’ webs but not their regularity is a function of the rate of growth and in consequence of the body dimensions.

Comparison of the FG and SG males’ webs at the same stage of maturation

All the FG webs photographed during the last stage were com- pared with all the SG webs photographed about ioo days later, during their last stage; this compares webs built at different times but comparable stages of maturity. Table 3 gives the figures for body weight, leg size, spiral area, thread length, me^h size and the variance of the mesh size. The webs of the FG (lighter) males had a spiral area larger, a significantly longer thread, smaller mesh size, and a higher regularity than the webs built by the SG males at comparable maturity (Figs. 8 and 9). The longer length of thread produced by the FG males than the SG males indicates that they have a better supply of silk or thinner thread. This may be sup- ported by the higher amount of food eaten per day and the higher rate of utilization of the food by the FG than SG males. The larger mesh size and irregularity of the SG webs is related to the larger body dimensions of these animals. The difference between the dimen- sions of the bodies of the FG and SG males coincides with a longer duration of development for the SG than for the FG males. We may assume that the regularity of spacing the spiral thread is related to the duration of the development in the two groups of males with different rate of growth, and is related to maturation within a group having a homogenous growth rate.

40

Psyche

[March-June

.s-s

M M

*s £ £ £ aT nj"

t: a

W) rt G »i cu b£ — o

-d o

C3 hG <U Oh

• - >

Oh C3 03 (jj

o g : .bC .M '

O «

fc rS

white lines being originally spaced 20 mm apart.

1973]

Ramousse — A raneus diadematus

4i

Mortality

The mean mortality in each group was :

set I set II

FG	1 6 1. 6 days post-hatching	146.5 days post-hatching
SG	3 dead (mean 329.0 days post-hatching) 3 still living (342.0 days post-hatching)	308-6 days post-hatching

The SG males lived significantly longer than the FG males (set I: T = 10, P — 0.01 ; set II: T gg 4, P = 0.05). Rapid growth occurs at the expense of endurance which is in agreement with the findings of Bonnet (1935), who found that the spiders’ lives short- ened with an increase in food supply, and the findings of Reed and Witt (1972), who found that the FG females of A raneus diade- matus lived shorter than the SG females.

In our laboratory, the males matured from July 1972 to January 1973. But in North America males of Araneus diadematus can only be found in September and early October in the Boston area, (Levi, 1971) and in Southern France in August and September (Bonnet, 1935). Nevertheless some authors found Araneus diadematus in the field during different seasons even in winter, Bertkau (1885) in Germany and Termeyer (1791) in Italy. Millot (1926) also ob- tained, in the laboratory, the survival of young Araneus diadematus during the winter; they completed their development the following spring. The biological cycle of Araneus diadematus varies according to the environmental conditions it goes through, and a low rate of feeding with the stable laboratory conditions could allow a longer lifespan for the spiders reared in the laboratory than in the field. In that case, the lengthening of the development merely emphasizes the difference betwen FG and SG animals.

The FG males survived an average of 80 days after the last molt in set I and 71.4 days in set II. These males grew and built webs approximately half their lives, then sought out mates. We can note equivalent facts for the SG males with their relatively long time scale.

42

Psyche

[March-June

General Discussion

The males grow until the last molt, at which time they attain their maximum weight; weight decreases slowly thereafter. The females have a distinctly different course of growth; their weight increased long after the last molt, generally until they lay a cocoon. The males mature more rapidly than the females, but the females grow bigger than the males and live longer.

The females and the males of each of the two sets of Araneus diadematus studied are clearly divided into FG and SG. The FG males are characterized by a higher frequency of building, rate of food consumption, rate of weight increase, rate of leg growth, rate of maturation as well as a smaller number of molts than the SG males, significant only for set I in the last instance.

The positive correlation between the rate of building and the rate of food consumption for all males must be expected, since the spiders were fed only when they had built a web. The different frequency of building may be explained as a lower threshold for web-building through hunger in the FG than in the SG spiders. The number of prey captured is a function of behavior mechanisms of the spider and potential prey; among the former are the stimuli that induce the spider to attack, the efficiency of this attack, and also a number of other variables such as web-site, web-characteristics, and frequency of building. The hunger stimulus which induces both, the attack and web-building, has a lower threshold for the FG than for the SG spiders, suggesting that in a natural habitat the FG males would be able to capture and eat more food than the SG spiders. In addition, the usual effect of genes on animals with rigid patterns is to alter behavior in a quantitative, rather than a qualitative, fashion (Mann- ing, 1967). The environmental conditions being the same for all the animals, the difference in threshold of hunger may be the conse- quence of different genotypes.

A relationship between the rate of food intake and the rate of growth indicates that the food was converted into spider tissues, in addition to maintain basal metabolism and to support the necessary activities like prey catching. The percentage of food converted into spider tissue was higher for the FG males than for the SG males, explaining the different growth rates. The same mechanism could provide a more ample supply of silk for the FG than for the SG males, which is suggested by the analysis of the web dimensions of

1973]

Ramousse — A raneus diadejnatus

43

the two groups. Hunger is an important drive for web-building and prey catching, which in turn increases the amount of food available to the spider. As a consequence, a good supply of food permits the spider to use more energy to metabolize tissues and silk, and a full supply of silk lowers the threshold of web-building. So, the fre- quency of building may be controlled by a changed feed-back between hunger and amount of food eaten.

A strong relationship exists between the rate of growth and the rate of maturation. But the number of molts was not constant, nor was the time separating two successive molts. The FG males went through fewer stages than the SG males (significant only in set I) and in less time. The rapid increase of the body weight of the FG spiders seems to force these spiders to change more often their rigid skins. Ecdysis is a crisis that requires extra energy to overcome; the heavy eating FG spiders could accumulate this extra energy in the form of reserves more rapidly than the SG spiders. The differential maturation may be attributed to nutrition. But nutrition is a func- tion of the amount of food eaten controlled through appetite, pro- ficiency in prey-catching, and web-building frequency. Similar rela- tions must explain the differences in development between males and females as well as between females of the same set.

Different schedules of feeding result in differential growth and maturation for spiders of the same set (Benforado and Kistler, 1972), suggesting that the amount of food eaten is a determinant factor. But with the same amount of food available, the spiders of the same set show different growth rates and maturation rates (Reed & Witt, 1972). This indicates that prenatal or genetic conditions control the development and maturation. In our study, the spiders could choose the food quantity they need through their behavior. When the spiders are in identical environmental conditions, we may assume that the difference in behavioral patterns present at hatching time probably are genetically determined. One pattern induces some spiders (FG) to capture and eat more food than other spiders, and in turn this large amount of food eaten by these spiders, increases their rate of development and maturation. The rapid growth in the two sets occurs at the expense of endurance and maybe weight in- crease, the FG males are short livers and small weighers. The short life-span of the FG spiders prevent them from mating with females of the same set, while some SG males live 'long enough to mate with the FG females of their own set.

44

Psyche

[March-June

Poetsch has shown (1963), that cocoons of a. single species of spider hatch at different times. This presumably provides an ad- vantageous distribution of egg-production over a period of time. The two cocoons studied hatched at different times, and the males of set II, which hatched fifteen days later, grew faster than the males of set I (significant only for the SG males). Between the sets the rapid growth occurs also at the expense of endurance and maybe weight increase. Differential growth occurs between the sets as well as within the sets, favoring the distribution of mature animals over a period of time. This suggests that during the favorable season mature males and females can mate and produce cocoons at various times, providing the species with a better chance to survive any drastic crisis due to the environment.

The relative quick maturation of males favors mating between animals of different sets and of different behavior instead of inbreed- ing. This allows the species to conserve a genetic pool with high selective potentialities, Dobshansky, 1951).

During development the last intermolt was distinct from the other stages. During this period the males built more webs, ate more food per day, and grew faster than during the other stages. The time separating the last two molts was generally longer than the time separating any other two successive molts. Sexual differentiation also took place during this period, and we may assume gametogenesis too. This would explain why the males need more food and a longer time to complete the last stage of development. The importance of the requirements during this time must make it the most difficult for the males.

Summary

The offsprings from two cocoons of Araneus diadematus , hatched at different times and placed in individual frames, were studied in the laboratory during the life-span of the males. During this time, the characteristics of the body (weight and size), the frequency and the parameters of the webs, the number and date of the molts, and the amount of food eaten were recorded for each animal. The spiders could choose their feeding schedules through their building behavior.

The males built and increased their weight only until the last molt, in contrast to the females which continued both building and increasing their weight long after the last molt. During the

1973]

Raiwousse — A rcineus diadematus

45

building period the males were distinctly different from the females only during the last stage. The males lived shorter and grew less than the females. The last intermolt was distinct from the other stages: the males built more webs, ate more food, grew faster than during the other stages.

Two different rates of development appeared among the males of each set, determining a fast and slow growing group. The fre- quency, the amount of food eaten, the rate of weight increase and the rate of maturation were higher for the fast growers than for the slow growers. As a consequence of the rapid growth, the life-span of the fast growing males was shorter and the maximum weight was lower (but not significantly) than for the slow growing males. Hunger and amount of food eaten determined the different growth rates and related maturation rates; a lower threshold is supposed for the fast growing males than for the slow growing males, and may be the consequence of a genetic difference. Maturation would be controlled by different patterns of behavior determined on a genetic level.

The differential maturation, which occurs within animals of a set and between sets, results in a distribution of mature animals over various times of year. The relative quick maturation prevents the fast growing males from mating with a female of the same set, but limited inbreeding is possible between the slow growing males and the females of the same set. A potential high survival of the species is assured by the dispersion of the individuals of one set and the dis- persion of the sets over different seasons.

ACKNOWLEDGEMENTS

This work was carried out in the laboratories of the Division of Research, North Carolina Department of Mental Health and was supported by Grant Number GB-25274 from the National Science Foundation to Dr. Peter N. Witt. The author gratefully ack- nowledges the assistance of Dr. Witt during all stages, the assistance of Mrs. Mabel Scarboro for all technical and laboratory work, Mrs. Rubenia Daniels for her administrative assistance, and of Dr. John O. Rawlings with whom the statistical tests used were discussed.

References cited Benforado, J. and Kistler, K. H.

1973. Growth of the orb weaver, Araneus diadematus, and correlation with web measurements. Psyche, 80: 90-100.

46

Psyche

[March-June

Bertkau, Ph.

1885. Ueber den Saisondimorphismus und einige andere Lebenser- scheinungen bei Spinnen. Zool. Anz. 8 : 459-464.

Bonnet, P.

1935. La longevite chez les Araignees. Bull. Soc. Etomol. de France. 40: 272-277.

Christiansen, A., Baum, R. and Witt, P. N.

1962. Changes in spider webs brought about by mescaline, psilocybin and an increase in body weight. J. Pharmac. ex. Ther. 136: 31-37.

Deevey, G. B.

1949. The developmental history of Latrodectus mactans (Fabr.) at different rates of feeding. Amer. Midi. Nat. Notre Dame, 42: 189-219.

Dobzhansky, T.

1951. Genetics and the origin of species. Columbia University Press.

Eberhard, W. G.

1971. The ecology of the web of Uloborus dwersus (Aranea: Ulobori- dae). Oecologia (Berlin), 6: 328-342.

Enders, R.

1972. Web site selection by Argiope aurantia Lucas and other orb weaving spiders (Araneidae). Thesis, N. C. State University, Raleigh.

Koenig, M.

1951. Beitrage zur Kenntnis des Netzbaus orbiteler Spinnen. Z. Tierp- sychol. 8 : 462-493.

LeGuelte, L.

1966. Structure de la Toile de Zygiella-x-notata Cl. (araignees, Agri- opidae) et quelques facteurs qui regissent le comportement de 1’araignee pendant la construction de la toile. These, Nancy.

Levi, H. W.

1971. The diadematus group of the orb-weaver genus Araneus north of Mexico (Araneae: Araneidae), Bull. Mus. Comp. Zool., 141 (4) : 131-179.

Manning, A.

1967. Genes and the evolution of insect behavior. Jerry Hirscb- McGraw-Hill (Behavior-genetic analysis).

Millot, J.

1926. Contribution a 1’ histophysiologie des Araneides. Bull. Biol. Fr. et Belg., Supp. 8: 1-238.

Peakall, D. B.

1964. Composition, function and glandular origin of the silk fibroions of the spider Araneus diadematus Cl. J. Exp. Zool., 156: 345-350.

1969. Silk synthesis, mechanism and location. Amer. Zoologist, 9: 71-79.

Peters, H. M.

1939. Uber das Kreuzspinnennetz und seine Probleme. Naturwissen- schaften 47: 776-786.

Potzsch, J.

1963. Von der Brutfiirsorge heimischer Spinnen. Wittenberg, Ziemsen.

Reed, C. F. and Witt, P. N.

1972. Growth rate and longevity in two species of orb-weavers. Bull. Brit. Archnol. Soc. 2(6) : 111-112.

1973]

Raiwousse — A rcineus diadematus

47

Reed, C. F., Witt, P. N. and Jones, R. L.

1965. The measuring function of the first legs of Araneus diadematus Cl. Behavior 25 : 98-119.

Reed, C. F., Witt, P. N. and Scarboro, M. B.

1969. The orb web during the life of Argiope aurantia (Lucas). Devel. Psychobiology 2(2): 120-129.

Reed, C. F., Witt, P. N., Scarboro, M. B. and Peakall, D. B.

1970. Experience and the orb- web. Devel. Psychobiology 3 (4): 251-265.

Sekiguchi, K.

1955. Differences in the spinning organs between male and female spiders. Sci. Rep. Tokyo Kyoiku Daigaku, 8: 23-32.

Termeyer, R. M. de

1791. Richerche e sperimenti sulla seta dei Ragni e sulla loro gen- erazioni. Scelte d’opusculi interessanti 3 : 288.

Turnbull, A. L.

1962. Quantitative studies of the food of Linyphia triangularis Cl. (Aranea: Linyphiidae) . Canad. Entomologist, 94(12) : 1233-1249.

White, C.

1952. The use of ranks in test significance for comparing two treat- ments. Biometrics, 8: 33-41.

Wiehle, J.

1927. Beitrage zur Kenntnis des Radnetzbaues der Epeiriden, Tetrag- nathiden and Uloboriden. Z. Morpholog. u. Okolog. der Tiere, 8 : 468-537.

Wolff, D. and Hempel, U.

1951. Versuche liber die Beeinflussung des Netzbaues von Zilla-x- notata durch Pervitin, Scopolamin and Strychnin. Z. vergl. Physiol., 33: 497-528.

Witt, P. N.

1963a. Interrelationships between web-building behavior and amount of thread material in the spider Araneus diadematus Cl. Proceed, of XVI Intern. Cong, of Zool.

1963b. Environment in relation to behavior of spiders. Arch of environ. Hlth., 7: 4-12.

Witt, P. N. and Reed, C. F.

1965. Spider web-building. Measurements of web geometry identifies components in a complex invertebrate behavior pattern. Science, 149: 1190-1197.

Witt, P. N., Reed, C. F. and Peakall, D. B.

1968. A spider’s web. Problems in regulatory biology. Springer-Verlag, New York.

Witt, P. N., Rawlings, J. O. and Reed, C. F.

1972. Ontogeny of web building behavior in two orb-weaving spiders. Am. Zoologist 12: 445-454.

ANNOTATIONS ON TWO SPECIES OF LINYPHIID SPIDERS

DESCRIBED BY THE LATE WILTON IVIE*

By P. J. van Helsdingen

Rijksmuseum van Natuurlijke Histone, Leiden, Netherlands

In December 1966 a short paper by Wilton Ivie was published in the Journal of the New York Entomological Society. In that paper the author described and illustrated two new species of Linyphiidae, Taranucnus durdenae from Pennsylvania, and Troglohyphantes ho-> koko, from Ontario. The latter was also recorded from the state of New York. In the course of my study of the North American Liny- phiidae it appeared that both were junior synonyms of North Amer- ican species, as will be discussed below. In a vast faunal region as North America such mistakes are quite understandable and not easily avoided, especially when a species originally was described in a rather ill-chosen genus, or when the older description of the species was based on the opposite sex.

While correcting the names of the two species, I take the oppor- tunity to discuss their distributions, habitats, and affinities wherever I have something of presumed importance to add. To avoid con- fusion the two species will be discussed separately under their correct names.

T aranucnus ornithes (Barrows)

Figures 1-10

Lepthyphantes ornithes Barrows, 1940, Ohio J. Sci., 40: 134, figs. 7-7C (descr. $ $ ; Ohio, Tennessee). — Vogel, 1967, Mem. Amer. Ent. Soc., 23: 93 (catal.) .

Taranucnus durdenae Ivie, 1966, J. New York Ent. Soc., 74: 224, figs. 1-5 (descr. $ $ ; Pennsylvania). — Vogel, 1967, Mem. Amer. Ent. Soc., 23: 100 (catal.). Vogel, 1968, J. New York Ent. Soc., 76: 102 (Penn- sylvania). NEW SYNONYMY.

Types. — Lectotype cf of Lepthyphantes ornithes , by present desig- nation, from Sugar Grove, Ohio. There are two $ paralectotypes, from Sugar Grove, Ohio, and the Great Smoky Mountains National Park, Tennessee. All three specimens are in the Barrows Collection at the Ohio State University, Columbus, Ohio, and were examined.

*Based on research done at the Museum of Comparative Zoology and published with a grant from the Museum of Comparative Zoology.

Manuscript received by the editor March 25, 1973

48

1973]

van Helsdingen — Linyphiid Spiders

49

Figure 1. Map showing distributions of Taranucnus ornithes (Barrows) (•) and Oreonetides recurvatus (Emerton) (★).

The cf holotype and $ paratype of Taranucnus durdenae, from Rec- tor, Pennsylvania, should be in the American Museum of Natural History, New York, but they were not examined by me.

In the Barrows collection, through kind cooperation of Dr. C. A. Triplehorn, one vial with Lepthyphantes ornithes was located. This vial contained one and two ? specimens, together with a label referring to the Ohio locality. A second vial, with the recorded ? from the Smoky Mountains, could not be found. It appears, that the single $ specimen from the latter locality has been added to the Ohio material. The original description mentions one $ from each locality (page 134, ninth line from bottom), while now two females are present in the Ohio vial. The description does not give any clue as to the sizes of the different specimens, and the mixed series there- fore cannot be separated again in accordance with the origin of the specimens.

Name. — I vie dedicated his new species to Beatrice Vogel Durden. Barrows did not give any explanation of the derivation of the name ornithes , but his remarks on the female epigyne gives us a key: “The epigynum (Fig. 7A) appears as if made up of two gasteropod shells with the large openings toward the abdomen. When seen from be- hind the two parts appear as two bird heads placed beak to beak/' The name therefore appears to refer to the two birds, revealed in a

50

Psyche

[March-June

posterior view of the epigyne (' ornis is Greek for bird, ornithes is the plural form). It is, of course, pure coincidence that durdenae, de- rived from the name Beatrice Vogel obtained through her marriage, has to be discarded in favor of the older ornithes, a name that still links the species with Beatrice Vogel, but now through the plural form of her maiden name (Vogel is the Dutch and German word for bird, ornithes is Greek for birds).

Distribution. — The number of available specimens has consider- ably increased during the years. The distribution (see map, Fig. i) now comprises, beside Pennsylvania, Tennessee and Ohio, also North Carolina, W. Virginia and Virginia. The species is probably not rare, but, in my own experience, is not easily collected because of its concealed habitat. Barrow’s Ohio specimens came from “under a log in a wooded ravine” (October), Ivies specimens from Pennsylvania were collected in July, and so was Vogel’s specimen, which came from the same locality in Pennsylvania. No other data, on the habitat are available from literature. I have collected a fair series (21? 2cJ ) in the Great Smoky Mountains National Park along the trail to the Alum Cave Bluffs. This trail leads off the road from Sugarlands to Newfound Gap in northern direction to Mount Le Conte, and is on the Tennessee side of the park, though close to the North Carolina line.

At the end of nearly three weeks of fruitless search for this species at both sides of the Smoky Mountains Range, we found a few speci- mens on what was planned to be our last day in this area. The next day we returned to the same spot and added a few more specimens to our collection. The species was found to inhabit small cavities and crevices in the steep rocky sides of the trail, and also in the dark hol- lows under and between the roots of trees. The forest at this height, between 1100 and 1300 m, has a heavy undergrowth of Rhododen- dron spec, and Dog-hobble (Leucothoe fontanesiana). The spiders were very difficult to collect. Even when we knew where to find them, many escaped by retreating from the entrances of their little caves into dark and impenetrable depths.

The distribution, as presented in Figure 1, is based on specimens examined by myself as well as on data supplied by my colleague and friend Dr. William A. Shear. The collecting dates of the various samples range from May until October. There is only one other mention of the exact circumstances of collecting: William Shear collected i? by sweeping tall weeds near Athens in West Virginia, a situation that does not agree with the habitat described above (which may have been delimited too rigidly).

1973]

van Helsdingen — Linyphiid Spiders

5i

Figures 2-6. Taranucnus ornithes (Barrows). Figs. 2-3, epigyne, ven- tral and dorsal aspect, respectively. Fig. 4, vulva, route to be followed by embolus indicated by consecutively numbered arrows. Fig. 5, tegulum and median apophysis ( ma ) of male palp. Fig. 6, median apophysis, dorsal aspect.

52

Psyche

[March-June

The genus Taranucnus is represented in Europe by the type-spe- cies, T. setosus (O.P.-Cambridge) , which I have collected from sev- eral localities in The Netherlands, and by T. bihari Fage, a troglo- biontic species from Rumania, which I do not know. A diagnosis of the genus, as based on setosus and \ornithes, reads as follows.

Small spiders (3.2 mm or less). Cephalothorax not much longer than wide. Eyes large, with PME closer to PLE than to each other. Chelicerae with stridulating files, dorsal margins with three teeth. Legs slender and very long (femur I ca. 2 times length cephalo- thorax, tibia I even slightly longer) . Legs spinose, including femora and metatarsi. Metatarsus IV without trichobothrium, Tm I 0.15- O.25. Abdomen with pattern, composed of blackish bars and areas but without white pigmented spots. Male palp with a short tibia, a rather flat and broadly rounded cymbium, which has a strongly modi- fied proximal part, and with a long, thin embolus; the embolus is supported by a well-developed embolic membrane. Epigyne with membraneous, coiled ducts leading to small receptacula; no socket or semi-covered depression visible for the reception of an apophysis of the male palp.

Both species prefer dark, protected places for a habitat: T. ornithes in crevices and cavities, T. setosus under overhanging vegetation (e.g. thick layers of heath) at the border of fens. The habitat of T. bihari would very well fit into this picture.

T aranucnus clearly fits into the tribe Linyphieae and seems very close to Labulla.

T aranucnus ornithes and T. setosus j beside their occurrence on dif- ferent continents, differ from each other mainly in the genitalia. Also ornithes seems to be the smaller species with more slender legs and a slightly more proximal Tm I.

The descriptions of T. ornithes by Barrows and Ivie can be sup- plied with the following data.

Measurements (in mm). Male, total length 2.0-2. 7, length cephalo- thorax 1. 05-1. 3, width 0.9-1.05. Female, total length 2. 2-3.0, length cephalothorax 1.05- 1.2, width 0.85-1.0.

Ratio length to width of cephalothorax ca. 0.85, which is high if compared with other, related genera (cf. Linyphia 0.65-0.75, Neriene °-55~o.75). Eyes large (diameter PME 0.09 mm) and close to- gether, the PME one-half diam. apart, and closer to each other than to PLE. AME not much smaller than PME. Chelicerae with dis- tinct stridulating files; three teeth in dorsal and three in ventral row.

Legs very long and slender: femur I 2 times as long as cephalo-

1973] van Helsdingen — Linyphiid Spiders 53

Figures 7-10. T aranucnus ornithes (Barrows), male palp. Fig. 7, lateral aspect. Fig. 8, radix (r) with embolus (e). Fig. 9, paracymbium (pc) and basal section of cymbium, showing modifications, lateral aspect. Fig. 10, cymbium, dorsal aspect.

54

Psyche

[March-June

thorax ( cf ) or slightly less ($), tibia I 22-24 ( 6 ) or 20-22 ($) diameters of segment long. Tibia I (without patella) slightly longer than femur, metatarsus shorter than tibia. Femora I-III with a d- spine on one-fourth of length, femur I with an additional l'-spine. All tibiae with the usual d-spines (basal d'^-spine1 at 0.3-0.35) ; tibia I with a pair of v-spines, a 1' and a l"-spine, and a whorl of apical spines; tibia II with a v" and a l"-spine only, III and IV with- out ventrals or laterals, but apical spines present on all tibiae. Meta- tarsi all with a single d-spine on O.20-0.25. All spines long (3 times diameter of segment or more) and thin. Tm I 0.16-0.20.

Abdomen as depicted by Ivie (1966: fig. 5), characterized by the inverted heart-shaped dark grey spot dorsally, followed by chevron and cross-bars of the same colour.

The malp palp was depicted and described at some length by Ivie (1966: 224, figs. 3-4). I add the following remarks to his observa- tions (Figs. 5-10). Patella and tibia short and simple, the patella bearing a strong dorsal spine, the tibia with a number of hardly thickened spine-hairs. Cymbium (Figs. 9-10) with complexly modi- fied basal parts, but with a simple, elbow-shaped paracymbium with the tip of the free arm widened into a flat, rounded plate; cymbium proper with two sclerites, a mesal and a lateral one, which are more sclerotized than the distal part of the element, and which enclose, together with the distal part, a deep dorsal depression of the cym- bium ; mesal sclerite distinctly connected without seam along the sclerotized mesal brim of the cymbium, lateral sclerite apparently with membraneous connections only. Median apophysis (Figs. 5-6, ma) with slender base but broadening into a flat and thin plate-like structure; no tooth or sharp tip; spermduct leaving element at the inside of the curvature. Radix (Fig. 8, r) a short, centrally situated element, spermduct running through the element and showing a dila- tion in the middle (Fickert’s gland?). Embolus (Fig 8, e) with firm base, long and tapering to a thin, thread-like tip. Embolic membrane a flat but twisted membraneous structure with narrowly upturned brim, arising from connecting membrane between median apophysis and radix.

Epigyne (Figs. 2-3) showing at either side a chitinous roof over the entrance of the duct, mesally separated by a narrow incision. Vulva (Fig. 4) showing the spirally coiled ducts, which run in loops in anterior direction but then turn backward again toward the thick-

JAs usual, the directions of the individual spines are noted by means of single or double accents, e.g. d' for pro-dorsal, d" for retro-dorsal, l' for pro-lateral, l" for retro-lateral, etc.

1973]

van Helsdingen — Linyphiid Spiders

55

walled portions close to the receptacula, the latter situated against the posterior wall at the outside of the entrances. Receptacula small.

It is clear that the embolus is supported by the embolic membrane in the unexpanded palp, -and it is not unlikely that it also guides and supports the long embolus during the difficult process of introducing the long and flexible element into the vulva. The whole palp may find firm support against the epigyne by means of the intricate dor- sal modification of the cymbium. The median apophysis does not show any hook-shaped parts (cf. Linyphia Neriene , etc.) and, at the most, may serve the purpose of supporting the functioning palp during copulation by being pressed with its broad and flattened apical portion against the epigyne. The embolus of T. ornithes does not have the small, toothed apophysis at the base of the embolus as shown in the figures of T. setosus by Merrett (1963: 382, fig. 39).

Oreonetides recurvatus (Emerton, 1913)

Figures 1, 11-17

Bathyphantes recurvatus Emerton, 1913, Trans. Connecticut Acad. Arts Sci., 18: 218, pi. 2 fig. 8 (descr. $, Vermont). — Ivie, 1969, Amer. Mus. Novit., 2364: 7 (= Oreonetides r.) .

Aigola recurvata ; Crosby, 1937, Proc. Biol. Soc. Washington, 50: 40, pi. 1 fig. 10 (descr. $, New York).

Troglohyphantes kokoko Ivie, 1966, J. New York Ent. Soc., 74: 226, figs. 6-7 (descr. $, Ontario; New York). NEW SYNONYMY.

Types. — cf holotype of Bathyphantes recurvatus from Gore Mountain, Norton, Vermont, in MCZ (examined). ? holotype and 9 paratype of Troglohyphantes kokoko from Ko-ko-ko Bay, Lake Timagami, Ontario, reported to be in the AMNH (not seen).

Additional material. — The specimens from Mt. Whiteface, New York, mentioned by Crosby (1937) have been examined in the AMNH. More recently collected specimens ( i9 3 cf ) come from George Lake, Alberta, Canada (CNC), 20.IX-1.X.1966 (Fig. 1).

Oreonetides recurvatus is a small species with long, slender legs, which are well provided with spines on femora, tibiae, and meta- tarsi. Tibia I has, beside the dorsal spines, a 1', a 1" and 2 v-spines. The chelicerae have three teeth in the dorsal row. The abdomen shows a dorsal pattern of grey cross-bars. Taking all together, it is not very likely that the present species belongs to Oreonetides , but it is maintained there for the time being. The genus is due for re- vision, and it does not seem appropriate to create a new genus for recurvatus here without having studied the other species presently residing in Oreonetides.

56

Psyche

[March-June

Figures 11-14. Oreonetides recurvatus (Emerton). Fig. 11, epigyne, dor- sal aspect. Figs. 12-13, vulva, ventral (12) and dorsal (13) aspects. Fig. 14, radix (r), embolus ( e ) and embolic membrane (em) of male palp, mesal aspect.

57

1973] van Helsdingen — Linyphiid Spiders

A first attempt to revise Oreonetides has been published recently (Saaristo, 1972). The paper contains a good characterization with excellent figures of the type-species, O. vaginatus (Thorell), but the many other, mostly Nearctic, species are not included. The diagnosis of the genus, therefore, might be too narrow to fit the other species, though it very well may be necessary to divide the genus into a num- ber of smaller units. The creation of Montitextrix by Denis (1963) for O. glacialis (L. Koch) was a first step in that direction. Oreone- tides flavus (Emerton) and O. rotundas (Emerton), both from the Nearctic region, are very close to M. glacialis in their genital struc- tures, but differ in the positions of the Tm I (0.65-0.70 in glacialis , 0.30-0.40 in flavus and rotundus) and the presence of a trichoboth- rium on metatarsus IV in glacialis (absent in all others).

A few species are more closely related to — i.e. more closely re- semble— the type-species vaginatus, viz., filicatus, firmus and abnor- mis, and possibly also rectangulatus. I do not see the principal dif- ference between vaginatus on the one hand, and fir?nus and abnormis on the other (cf. Saaristo, 1972: 70). Oreonetides might constitute an example of transition from the still folded scapes (but how un- der-developed in comparison with the flexible and elaborately built scapes of Lepthyphantes species!) of vaginatus and filicatus to the rigid and unfolded scape of firmus and abnormis, where the narrow portion of the scape with the socket or semi-covered depression is not present. The absence of a well-developed median apophysis (Saaristo: suprategulum ) is, of course, correlated with this simple build of the epigyne, and should not be used as an independent character.

Oreonetides filicatus is a good species and not a synonym of vaginatus as suggested by Saaristo (1972: 72). It is smaller than vaginatus, but both male palp and epigyne are resembling each other, though they differ in detail. Without question the species is con- generic with vaginatus. However, the anterior tibia does not bear the T-spine, which is one of the characters mentioned in the diagnosis of the genus by Saaristo (1972: 69). I put this forward here as a demonstration of my above remark on the too narrow delimination of the genus.

Oreonetides rectangulatus (Emerton), of which I only know the male, is different in several respects. Most striking are the pecul- iarly shaped chelicerae, which bear a strong conical protrusion on their dorsal surface for three-fifths of their length. The palp also deviates from the true Oreonetides type.

Oreonetides flavescens (Crosby), described in Aigola (Crosby, 1937: 39, figs. 7-8) from New York I have not seen, but judging

58

Psyche

[March-June

17

0-5 mm

Figures 15-17. Oreonetides recurvatus (Emerton), male palp. Fig. 15, lateral aspect. Fig. 16, cymbium, dorsal aspect. Fig. 17, tegulum with me- dian apophysis (ma) .

1973]

van Helsdingen — Linyphiid Spiders

59

from the figures and description it was correctly placed with vagina- tus and related species.

To the descriptions of O. recurvatus given by the various authors

I add the following remarks.

Measurements (in mm). Male, total length 2. 7-3.0, length cephalo- thorax 1.25- 1.43, width 1.05- 1.20. Female, total length 2.7 (Ivie, 1966: 2.8 mm), length cephalothorax 1.22 (1.3), width 1.05 (1.1).

Ratio length to width of cephalothorax 0.85. PME large (0.09 mm), separated by their diameter, same distance to AME which are barely smaller; lateral eyes of posterior and anterior rows con- tiguous. Chelicerae showing sexual dimorphism; c? with three “teeth” on dorsal margin: a basal, round tubercle with a more slender and sharp tooth just dorsally of its tip, a second tooth at some distance of the basal tubercle, and with a small tubercle at the end of the row; ventral row with two small teeth; 9 also with three dorsal “teeth”, ventral row with five small teeth (three ventral teeth ac- cording to Ivie, 1966). Stridulating files well-developed, ridges fine and close together.

Legs long and slender, femur I 1.8- 1.9 times length cephalothorax, tibia I 20-23 ( c? ) or 18 (9) diams. of segment long. Tibia I longer than femur I, metatarsus as long as femur ( cf ) or slightly shorter (9). Femora I-III with a d-spine (on 0.5), femur I with an addi- tional l'-spine. All tibiae with 2 d-spines (position basalmost d- spine 0.25-0.35), tibia I also with one 1', one 1", and 2 v-spines, tibia

II with one v-spine and a 1 "-spine; apical spines on segments hardly developed; all metatarsi with a single d-spine (position ca. 0.20 ). Tm I 0.15-0.20, metatarsus IV without trichobothrium.

Male palp (Figs. 14-17). Characteristic is the meso-proximal hook-like projection of the cymbium, which points laterad (Fig. 16). The paracymbium lacks the isolated ventral hair which occurs in O. vaginatus (see Saaristo, 1972), and also in O. filicatus and rotundas; only a small tubercle present on this spot. Median apophysis (Fig. 17, ma) well-developed, with slightly curved tip. Embolic section (Fig. 14) in general structure resembling vaginatus , but without lamella; ventro-lateral branch rounded-truncated, a slender, chiti- nous projection present between this branch and base of embolus, and a second, equally narrow, slightly larger, membraneous projection, arising from the dorsal surface of the radix. Embolus {e) short and squat, curved, with a sharp spermduct tooth, and well-protected by the embolic membrane {em) , which arises from the dorsal surface of the radix from the membraneous connection between median apophy- sis and radix (r).

6o

Psyche

[March-June

Epigyne and vulva (Figs. 11-13). Epigyne as depicted by Ivie (1966: figs. 6-7). Scape folded and reappearing from below the main body of the scape with a rounded, membraneous tongue, which possesses a semi-covered depression. Entrances of ducts situated in the main body of the epigyne, laterally and behind the bend of the scape; ducts converging in forward direction, then curving outwards to turning-points and to the receptacula seminis, which lie close to the turning-points. The short fertilization-ducts curving to dorsal side and ending as open gutters.

From the structures of palp and epigyne it is clear that the me- dian apophysis, in connection with the depression at the tip of the scape, serves as an important means of support during copulation. The scape of the epigyne looks rather rigid and probably cannot be pulled out of its resting position very far (cf. Lepthyphantes) , though the ventralmost part may get pushed away from the main body so as to allow the embolus to reach the entrance of the duct of the epigyne. The exact functions of the hook-like projection of the cymbium, the proximal, roughened extension of the paracymbium, and the lateral arm of the radix are not easily understood without the aid of actual observation of the pairing in one of the species of this group.

Acknowledgements

The present study is part of a general survey of North American Linyphiidae, supported by Public Health Service Research Grant AI-01944, from the National Institute of Allergy and Infectious Diseases, to H. W. Levi.

The help with types and other specimens by the following institu- tions and their curators is thankfully acknowledged : Dr. Herbert W. Levi, Museum of Comparative Zoology, Harvard University, Cambridge, U.S.A. (MCZ) ; Dr. John A. L. Cooke, American Museum of Natural History, New York (AMNH); Dr. C. A. Triplehorn, Ohio State University, Columbus, Ohio; Dr. Robin E. Leech, Canadian National Collection, Ottawa, Canada (CNC). Thanks are also due to Dr. William A. Shear, Concord College, Athens, W. Virginia, for information on his collection.

References

Barrows, W. M.

1940. New and rare spiders from the Great Smoky Mountains National Park region. Ohio J. Sci., 40: 130-138, figs. 1-12.

Crosby, C. R.

1937. Studies in American spiders: the genus Aigola Chamberlin. Proc. Biol. Soc. Washington, 50: 35-42, pi. 1.

1973]

van Helsdingen — Linyphiid Spiders

61

Denis, J.

1963. Araignees des Dolomites. Atti 1st. Veneto Sci. Lett. Arti, 121: 253-271, figs. 1-16.

Emerton, J. H.

1913. New England spiders identified since 1910. Trans. Connecticut Acad. Arts Sci., 18: 209-224, pis. 1-2.

I VIE, W.

1966. Two new North American spiders (Araneae: Linyphiidae). J. New York Ent. Soc., 74: 224-227, figs. 1-7.

1969. North American spiders of the genus Bathyphantes (Araneae, Linyphiidae). Amer. Mus. Novit., 2364: 1-70, figs. 1-121. Merrett, P.

1963. The palpus of male spiders of the family Linyphiidae. Proc. Zool. Soc. London, 140: 347-467, figs. 1-127.

Saaristo, M. I.

1972. Redelimitation of the genus Oreonetides Strand, 1901 (Araneae, Linyphiidae) based on an analysis of the genital organs. Ann. Zool. Fennici, 9: 69-74, figs. 1-17.

Vogel, B. R.

1967. A list of new North American spiders (1940-1966). Mem. Amer. Ent. Soc., 23 : 1-186.

1968. Additional records of spiders from Western Pennsylvania. J. New York Ent. Soc., 76: 101-105.

CORRELATION BETWEEN SEGMENT LENGTH AND SPINE COUNTS IN TWO SPIDER SPECIES OF ARANEUS (ARANEAE: ARANEIDAE)*

By L. David Carmichael Museum of Comparative Zoology

Abstract

Observations made on several hundred adult male spiders of two species of Araneus indicate a highly significant correlation (p<0.ooi) between the length of a segment (tibia of the second leg) and the number of macrosetae (“spines”) present on the seg- ment. This result is further supported by observations on the first tibiae of about twenty male A. trifolium, one of the two species, and by a few observations on immatures of the two species. A short summary of the methods used in taking the measurements and mak- ing the calculations is followed by discussion of the implications of this correlation with reference to species determination and geo- graphic variation.

Methods

The study was done on two species of common North American spiders, Araneus trifolium (Hentz) and A . marmoreus Clerck. 185 male specimens of A. trifolium yielded 347 tibiae of the second leg (some specimens had lost one leg) ; the length of this segment was measured, and the number of spines on the segment was counted. In addition, lengths and spine counts were taken for the first tibiae of 23 of the spiders, yielding 41 observations. Similarly, 120 speci- mens of A. marmoreus yielded 210 second tibiae; the length was measured, and two spine counts were made: the total number on the tibia, and the number of modified, “dentiform”1 spines (see Figure 3). The samples of both species were museum collections, and represented almost the entire known range of each in North America, extending from coast to coast and roughly from the 35th to the 55th parallels.

The spines of the second tibia, like all the spines of these spiders, are actually setae in the entomological sense ; they are set in a socket

* Manuscript received by the editor January 10, 1973

Term used by Locket & Millidge (1953), pp. 120, 121.

62

1973]

Carmichael — A raneus

63

Figures 1 and 2. A. trifolium tibia II, right leg. 1, anterior (prolat- eral) surface; 2, posterior (retrolateral) surface.

Figures 3 and 4. A. marmoreus tibia II, right leg. The dentiform spines are blackened. 3, anterior (prolateral) surface; 4, posterior (retrolateral) surface.

Note the difference in scale between 1, 2 and 3, 4.

64

Psyche

[March-June

of the cuticle, and are movable. Thus even when the spine itself becomes detached and lost, as is common with preserved specimens, its presence or absence can be determined unequivocally by the pres- ence or absence of the setal socket.

In the two species examined, the spines are uniformly larger than the hairs which are also present, the former having a diameter at the base of roughly 0.03 to 0.06 mm, while the latter are five to ten times smaller. In A . marmoreus, there is a second type of spine, described as dentiform, which is roughly 0.07 to 0.10 mm at the base. Since this constituted a distinct group, it was counted sep- arately.

Finally, the spines in both species are arranged in fairly constant patterns, particular to the species (see Figs. 1-4). This makes it possible to recognize each spine, which further eliminates any un- certainty as to the number of spines. These three factors, the clear difference between spines and hairs, the presence of a socket whether or not the spine itself has been lost, and the possibility of recogniz- ing each spine, make the spine counts unambiguous and as accurate as possible within the limits of observor error.

The length of the tibia was measured along the dorsal midline of the segment, between points “a” and “b” as shown in Figures 1 and 3. A grid in the microscope eyepiece, each cell of which meas- ured 0.325 mm on a side (as determined with a stage micrometer), permitted accuracy to ± 0.02 mm, or about =+= 1%.

The treatment of the data followed standard statistics texts; the actual calculations were performed by the Harvard SDS 940 digital computer.

Results

Table I presents the correlation coefficient for the number of spines versus the segment length in the four different samples, as well as the coefficient of regression (b) and the results of the t-test for b = o. These values indicate that in all four cases the correla- tion is highly significant (i.e. significant at the 0.1% level), b, the coefficient of regression, quantifies the relationship established here; it is the slope of the estimated regression line drawn in Graphs I and II. The line is:

(number of spines) = a + b X (length of tibia)

Table II gives the mean and variance for the two variables in both species, and the mean absolute difference between right and left legs of individual specimens. In general the results are very

1973]

Carmichael — A raneus

65

Graph I. Scatter diagram and estimated regression line for A. trifolium. measurements, b = 2.75, a — 15.24, (expected number of spines) — a + b X (tibia length). Open circle, adult specimens; closed circle, immature specimens; triangles, mean spine counts for given tibia! length.

66

Psyche

[March-June

similar to those found by Beatty (1967) for Adriana : the spine count is quite constant within each species, though few specimens are actually identical. Furthermore, in the two species of Araneus j as in Ariadna , almost no individuals are completely symmetrical in pattern, and most are asymmetric in actual spine counts. Beatty attributed such differences within and between individuals to develop- mental “accidents”; it is clear, however, that in the case of Araneus some of the variation between individuals is specifically related to difference in size. But for any single specimen the difference in spine count (between left and right legs) seems not to be correlated with the difference in segment length. (This correlation is cal- culated as r2 in Table in II; the values, though positive, are not sig- nificant at the 5% level.) Thus Beatty’s assertion is correct for individual spiders; differences between left and right legs do seem to be due to developmental accidents, and quite independent of each other. This point will be important in the following discussion.

Discussion

It is important first to note that the above correlation does not in itself imply cause and effect; this is clear from the fact that for any individual, segment length and spine count are not correlated. It is likely that both the length of the segment and the number of spines are dependent on some other factor, such as general body size, etc.

One obvious possibility is that both measurements are related to the degree of development, that is, to the number of molts the spider has undergone. In most spiders raised there is some variation in the number of preadult instars within a species. Furthermore, it is not known with certainty at what stage these spiders mature or how many molts occur after maturation, so this possibility cannot be ex- amined with the data available. All the calculations here are based on sexually mature specimens, but their ages cannot be determined more precisely. Consequently, part of the spine count variation may be dependent on this unmeasured variable; of course, size is some- what dependent on this variable too.

On the basis of the data presented here, the best statement is simply that spine count is very significantly correlated with segment length, in these two species of Araneus.

Then there is the question of geographic variation. The samples studied represent a pool of many local populations in North America, and it is possible that the relationship between segment length and spine count is different in different regions. (Preliminary examina- tion of the data with regard to this question indicate that this is in

1973]

Carmichael — A raneus

67

Graph II. Scatter diagram and estimated regression lines for A. mar- moreus measurements. Upper part of graph (open circles) shows total spine counts: b = 4.43, a = 16.52. Lower part of graph (closed circles) shows dentiform spine counts only: b=1.44, a = 8.12. Triangles show mean spine counts for given tibial length.

68

Psyche

[March-June

TABLE I

A. trifolium

II

Tibiae 347 0.54

2.75 ±0.60 12.0

Tibiae I 41 0.61

2-39 ± 1.36 4.8

A Total count 210 0.63

4.43 ± 1.00 11.6

marmoreus

Dentiform only 210 0.42

1.44± 0.56 6.7

N n b

t-test (for b — 0)

Calculation of correlation and regression coefficients for spine counts versus lengths of first and second tibiae of A. trifolium and for total and dentiform spine counts versus length of second tibia of A. marmoreus. ri is the correlation coefficient for spine counts versus length ; b is the slope of the regression line, presented with its 99% confidence intervals. The significance of the correlation may be found either from the value of the coefficient r or from the t-test on the null-hypothesis b — 0.

fact the case.) While this does not affect the validity of the results as they have been presented, it would modify the quantitative rela- tionship (expressed by b) significantly in separated areas. This ques- tion is open to further study; its significance will be mentioned below.

Conclusion

The correlation between the number of spines on a segment and the length of the segment is important to at least two aspects of Araneology: taxonomy and the study of geographic variation. Mac- rosetal counts have often been used to distinguish between different genera of spiders2, as well as between species of one genus such as Araneus. If the situation described in this paper is a general one, then clearly any character based on setal counts should be used for taxonomic purposes only after careful study. In general, it would seem from observations on these two species of Araneus that the number of spines alone is not highly reliable, but the pattern is quite constant within a species (or at least recognizable, though spines may be missing, or present in “unusual” locations). This is sup- ported by observations made on species of the genus Neoscona (Ber- man and Levi (1971), p. 467 ) .

Secondly, in studying geographic variation, it is necessary at least in this case to consider the mean dimensions of local populations as well as the spine counts. A marked variation in spine count between

2For examples, see Kaston (1948), with reference to: Gnaphosidae

(Drassodidae) pp. 347, 354; Clubionidae, pp. 367, 382; Thomisidae, pp. 410, 440; and Salticidae, p. 445.

1973]

Carmichael — A raneus

69

TABLE II

A. trifolium (N=166) A, marmoreus (N=98)

	tibial	spine	tibial	total	dentiform
	length (mm)	count	length (mm)	spine count	spine count
mean	2.31 ± 0.06	21.59 ± 0.31	3.39 ± 0.09	3 1. 54 ± 0.66	13.01 ± 0.32
standard
deviation	0.44	2.22	0.52	3.66	1.76
mean
difference (absolute)	*	1.25 ±0.23		1.82 ±0.43
r2	0.134	0.127	. —

Self-explanatory: the means are presented with their 99% confidence

intervals. See text for details.

two regions might be obscured by the fact that specimens from one of the regions are generally smaller than those from the other (which itself might be due to significant geographic variation in size, or to differences in sampling techniques, etc.).

Acknowledgments

I wish to thank Dr. H. W. Levi, my advisor at the time this study was done, for his advice and patient help; Dr. S. J. Gould, who advised me on the interpretation of statistical data; and the late Mr. Ivie of the American Museum of Natural History, who loaned me specimens from that museum. While doing the research for this paper, I was supported until June, 1969 by a scholarship from the National Merit Foundation, and subsequently by an NDEA title IV fellowship.

References

Beatty, J. A.

1967. The Spider Genus Ariadna in the Americas. Doctoral Thesis, Harvard University, Dept, of Biology.

Berman, J. D. and H. W. Levi

1971. The Orb Weaver Genus Neoscona in North America (Araneae: Araneidae). Bull. Mus. Comp, Zoo-1. 141 (8) : 465-500.

Kaston, B. J.

1948. Spiders of Connecticut. Bull. Conn. Geol. Natur. Surv. Vol. 70 : 1-874.

Locket, G. H., and A. F. Millidge

1953. British Spiders, 2. Ray Society (London).

Simpson, G. G., A. Roe, and R. C. Lewontin

1960. Quantitative Zoology (2nd ed.). Harcourt, Brace & World, Inc. Wetherill, G. B,

1967. Elementary Statistical Methods. Methuen & Co.

ANT LARVAE OF FOUR TRIBES:

SECOND SUPPLEMENT

(HYMENOPTERA: FORMICIDAE: MYRMICINAE)*

By George C. Wheeler and Jeanette Wheeler Laboratory of Desert Biology Desert Research Institute University of Nevada System Reno, Nevada 89507

Subsequent to the publication of our first supplement on the larvae of the subfamily Myrmicinae (1960a)1 we have received from other myrmecologists so much additional material that it has become neces- sary to publish additional supplements.

Tribe Leptothoracini Genus Macromischa Roger Machromischa subditivci Wheeler

Creighton 1965 — Life cycle: egg 30 days, larva 2 3 days, pupa 19 days.

Genus Leptothorax Mayr2

Kempf 1959: 393 — “The morphological distinctness of the im-

aginal stages and the distribution of the species may even suggest to accord Nesomyrmex full generic status. The larvae, however, are quite close to the holarctic subgenus Leptothorax s. str., according to G. C. & J. Wheeler (1955), who studied those of echinatinodis’1

Leptothorax carinatus Cole

semipupa: Length (through spiracles) about 2.2 mm. Profile

probably similar to L. amhiguus (1955: 22), otherwise differing in the following details. Body hairs (1) about 0.006 mm long; (2) 0.006-0.087 mm long; (3) about 0.14 mm long, four on the dorsum of each AI-AIII. Cranium transversely subelliptical. Head hairs 0.012-0.03 mm long, simple or bifid. Ventral border of each lobe of labrum with one isolated and three contiguous sensilla and a few minute spinules. Each labial palp a cluster of five sensilla. (Mate-

*Manuscript received by the editor January 30, 1973

To save space we cite our own papers by year and page; the complete references are in Literature Cited.

2In 1950, M. R. Smith changed the well established subgeneric names Leptothorax and Mychothorax to Myrafant and Leptothorax respectively. Could more confusion be generated in less than two pages? The established names should have been conserved. We refuse to accept these changes.

70

1973]

Wheeler Wheeler — Ant Larvae

71

rial studied: three semipupae from Texas, courtesy of Dr. A. C. Cole.)

Leptothorax hispidus Cole

worker semipupa. Length (through spiracles) about 3 mm. Similar to L. ambiguus (1955: 22) except as follows. No spinules on integument. Body hairs: (1) 0.013-0.025 mm long; (2) 0.025- 0.068 mm long; (3) about 0.2 mm long, four each on AI-AIIL Head hairs 0.01 5-0.038 mm long, with tip bifid. Labrum with about ten hairs, about 0.025 mm long, on the anterior surface; posterior surface with eight sensilla. Each maxillary palp a raised cluster of four sensilla; each galea a very short peg with two sensilla. Each labial palp with five sensilla.

young sexual larva. Length (through spiracles) about 3.2 mm. Similar to the above larva except as follows. Body sac-like. Dorsal surface of posterior somites with a few minute spinules. Cranium transversely subelliptical. Anterior surface of labrum with four hairs.

Material studied: four worker semipupae and two young sexual

larvae from Texas, courtesy of Dr. A. C. Cole.

Leptothorax nevadensis Wheeler

Length (through spiracles) about 2.8 mm. Similar to L. ambiguus (1955: 22) except as follows. Integument with a few minute spi- nules on the venter of anterior somites and the dorsa of posterior somites. Body hairs: (2) 0.025-0.19 mm long; (3) about 0.25 mm long, on AI-AIV. Head hairs 0.013-0.03 mm long, generally dis- tributed. Each mandible with narrow blade and two medial teeth. Each maxilla with a ventral projection on the lateral surface; each palp a cluster of five sensilla. Each labial palp a cluster of five sen- silla. (Material studied: six larvae from Oregon, G. C. and J. Wheeler #8)

Leptothorax niger fplendidiceps Urbani

Urbani (1968: 460-464) described the larva and figured young and mature larvae and the head of the latter.

Leptothorax nitens Emery

Length (through spiracles) about 2.6 mm. Similar to L. ambiguus (1955: 22) except as follows. Thorax slightly more constricted and arched ventrally. Integument of venter of anterior somites and dorsa of posterior somites with a few minute spinules, isolated or in short rows. Body hairs: (1) 0.006-0.012 mm long; (2) 0.025-0.125 mm long; (3) about 0.2 mm long. Head hairs 0.013-0.05 mm long,

72

Psyche

[March-June

Fig. 1. Leptothorax (M.) provancheri. a, larva in side view, Xl7; b, head in anterior view, X67; c, very young larva in side view, Xl7; d, simple and branched body hairs, X169; e, surface view of dendritically branched body hair, X169; f, anchor-tipped body hair, Xl69; g, left mandible in anterior view, X163. Fig. 2. Rogeria procera . a, left mandible in anterior view, X169; b, head in anterior view, X 67 ; c and d, branched body hairs, X264; e, anchor-tipped body hair, X264; f, larva in side view,, X1B*

1973]

Wheeler & Wheeler — Ant Larvae

73

simple or bifid. Labrum with eight hairs on the anterior surface; posterior surface with eight sensilla. Each mandible with the blade narrow and bearing two teeth. Each maxillary palp a cluster of four sensilla. Each labial palp a cluster of five sensilla.

very young larva. Length (through spiracles) about 1.3 mm. Similar to very young larva of L. ambiguus (1955: 23) in shape, otherwise similar to mature larva above except as follows. Cranium more rounded. Head hairs 0.013-0.038 mm long, all simple. Labrum more narrowed ventrally. Mandibles with narrower blade and sharper teeth. Maxillae very small ; each galea a slight elevation with two sensilla.

Material studied: 15 larvae from Oregon, G. C. and J. Wheeler

# 14.

Leptothorax (Mychothorax) provancheri Emery (Fig. 1)

Length (through spiracles) about 3.7 mm. Paraponeriform (i.e., shaped somewhat like a crookneck squash ; neck short and stout ; body elongate, stouter, straight and subcylindrical) ; posterior end rounded; a ventrally projecting boss on each lateral surface of T r; each thoracic somite and AI-AIII with a hairless midventral boss. Anus postero- ventral. Leg, wing and gonopod vestiges present. Diameter of spira- cles decreasing posteriorly. Integument of posterior somites with minute spinules in short rows dorsally and isolated ventrally. Body hairs rather sparse. Of three types: (1) 0.025-0.45 mm long, with straight or kinked shaft and branched (bifid to dendritic) all branches with denticulate tip, on all surfaces of all somites except venter of Ti; (2) 0.025-0.1 mm long, simple, on all surfaces of Ti, fewer on T2 and T3; (3) 0.25-0.375 mm long, anchor-tipped, with curled to kinked shaft, four in a row across the dorsum of each AI-AV (sometimes also one on AVI). Cranium subhexagonal, longer than broad ; sides of head nearly straight. Each antenna on a teardrop- shaped base; each a slight dome with three sensilla, each of which bears a spinule. Head hairs numerous, minute (0.003-0.019 mm long). Labrum paraboloidal in anterior view; anterior surface with 12 hairs (about 0.012 mm long), and two sensilla; ventral border with two isolated and two clusters of three sensilla each; posterior surface with about eight sensilla and a few minute spinules in short rows. Mandibles leptothoraciform (i.e., moderately narrow; taper- ing gradually and curving gradually to an apical tooth; anterior surface produced medially into a blade bearing two subapical teeth) ; all teeth subequal. Each maxilla with the apex conoidal ; each palp

74

Psyche

[March-June

a short skewed peg with five sensilla; each galea a short frustum with two apical sensilla. Labium narrowly paraboloidal ; sparsely spinulose, the spinules minute and in short transverse rows ; each palp represented by a cluster of five sensilla; an isolated sensillum be- tween each palp and the opening of the sericteries, the latter a short transverse slit. No spinules on hypopharynx.

very young larva. Length (through spiracles) about 1.6 mm. Abdomen sac-like, thorax forming a stout neck. Integument of ven- ter of anterior somites and entire surface of posterior somites with minute spinules. Body hairs sparse. Of three types: (i) o.oi 3-0.1 mm long, on dorsal and lateral surfaces of thorax and on dorsa of AI-AIV; (2) about 0.006 mm long, on venter of Ti, few on AV : (3) about 0.1 mm long, four each on AI-AIV. Antennae minute, each with three sensilla. Head hairs about 1/3 as numerous, minute (0.002-0.005 mm long). Labium subrectangular ; anterior surface with about ten sensilla. Mandibles subtriangular in anterior view, with all teeth sharp-pointed. Each maxillary palp represented by a cluster of five sensilla; each galea represented by two contiguous sensilla. Otherwise as in the mature larva.

Material studied : numerous larvae from Colorado, G. C. and

J. Wheeler #16.

The larva of L. provancheri resembles our other species of Mycho- thorax in profile and in mandible shape, but it differs markedly in the shape of the dominant type of body hair, in the shape of the head, in the shape of the labrum, and in the abundance and size of head hairs.

Genus Rogeria Emery

Because we had inadequate material previously (1955: 28) we are giving a complete description below.

revised description. Profile solenopsidiform. Body hairs mod- erately abundant ; of two types — ( 1 ) short, generally distributed and variously branched and (2) anchor-tipped, on mesothorax, meta- thorax and first three abdominal somites. Head hairs moderately numerous, moderately long and bifid. Mandibles leptothoraciform.

In our 1960b key Rogeria would go to “ Monomorium antarcti- cum” ( = Chelaner) , from which it cannot be distinguished generic- ally.

Rogeria pro c era Emery (Fig. 2)

Length (through spiracles) about 4 mm. Solenopsidiform (i.e., short and stout; head ventral, near the anterior end; prothorax bent

1973]

Wheeler & Wheeler — Ant Larvae

75

ventrally to form a very short stout neck ; remainder of body straight ; both ends broadly rounded. Anus ventral.) Dorsal profile long and C-shaped; ventral feebly sigmoid. Leg vestiges present. Spiracles small; diameter diminishing posteriorly. Integument of venter of thorax with relatively coarse spinules in transverse rows ; a few min- ute spinules on dorsa of posterior somites. Body hairs moderately abundant. Of two types: (i) 0.044-0.138 mm long, with tip short- bifid to long-branched, the branches variously denticulate or branched ; (2) about 0.25 mm long, anchor-tipped with flexuous shaft, six on the dorsum of each T2, T3, AI, All and four on AIII. Cranium subtrapezoidal, broadest dorsally; occiput nearly flat; clypeus bulg- ing. Antennae each with three sensilla, each of which bears a rather long spinule. Head hairs moderately numerous, 0.05-0.075 mm long, bifid with the branches short to long. Lab rum bilobed, narrowed dorsally ; each lobe with two hairs on the anterior surface about 0.006 mm long, ventral border with three isolated and two contiguous sensilla, posterior surface with six isolated and a cluster of three sensilla; entire posterior surface with a few short rows of minute spinules dorsally and with coarse isolated spinules ventrally. Mandi- bles leptothoraciform (i.e., moderately stout, tapering gradually and curving gradually to an apical tooth, anterior surface produced me- dially into a blade, which bears two medial teeth and a few denticles). Each maxilla paraboloidal, apex with coarse isolated spinules; each palp a cylinder with four apical and one subapical sensilla; each galea a short stout cylinder with two apical sensilla. Labium nar- row, anterior surface with coarse spinules, which are isolated or in short rows near each lateral surface; each palp a slight elevation with five sensilla; an isolated sensillum between each palp and the opening of the sericteries, the latter a transverse slit. (Material studied: 12 larvae from Brazil, courtesy of Drs. W. L. Brown and K. Lenko.)

Tribe Ocymyrmecini Genus Ocymyrmex Emery

Profile aphaenogastriform. Prothorax narrowed rapidly to the diameter of the head. Head small. Anus with a prominent poste- rior lip. Body hairs numerous, short, with frayed tip. Cranium subcircular. Antennae high on cranium, minute and in pits. Head hairs few, short, with bifid tip. Labrum paraboloidal, as long as broad. Mandibles vollenhoviform but with only one medial tooth.

In our 1960b key this genus would fit under Group D but would require a new rubric: 6. Body aphaenogastriform (Di) ; mandibles vollenhoviform (Ilf).

76

Psyche

[March-J une

Ocymyrmex arnoldi Forel

(Fig. 3)

Length (through spiracles) about 5.9 mm. Aphaenogastriform (i.e., stout and rather elongate; diameter greatest at AIV and AV ; slightly constricted at AI ; thorax stout and arched ventrally, but not differentiated into a neck; posterior end broadly rounded, anus ventral). Prothorax narrowed rapidly to diameter of head. Head small. Anus with a prominent posterior lip. Leg and wing vestiges present. About six differentiated somites. Spiracles small, dimin- ishing slightly posteriorly. Entire integument spinulose, the spinules minute and in short transverse rows, the rows longer and closer together on the venter of the anterior somites. Body hairs abundant, uniformly distributed, all short (0.025-0.075 mm long), with stout shaft and frayed tip. Cranium subhexagonal ; mouth parts rather large. Antennae minute, each in a small pit bounded medially by a high rim, three sensilla each bearing a tall spinule. Head hairs few, short (0.019-0.038 mm long), slightly curved, with short-bifid tip. Labrum paraboloidal, with three small ventral projections; anterior surface with six minute hairs; ventral border with two small sensilla on each ventrolateral projection and two groups of two larger con- tiguous sensilla ; posterior surface with 16 sensilla and scattered min- ute spinules. Each mandible vollenhoviform, but with only one medial tooth (i.e., slender, rather long and nearly straight, apex form- ing a moderately long slender tooth which is slightly curved medially ; with a narrow medial blade, from the edge of which arises one in- conspicuous medial tooth). Each maxilla with the apex paraboloidal and with minute isolated spinules; each palp a narrow frustum with four apical and one lateral sensilla ; galea digitiform with two apical sensilla. Labium paraboloidal, with a few short rows of minute spinules; each palp a skewed peg with five sensilla; an isolated sensil- lum between each palp and the opening of the sericteries, the latter a transverse slit. Hypopharynx with minute spinules in short rows. (Material studied: three larvae and one semipupa from Rhodesia, courtesy of Dr. W. L. Brown.)

Trire Tetramoriini Genus Tetramorium Mayr T etramorium caespitum (Linnaeus)

Bruder and Gupta 1972 — Description p. 366; photographs of first, second and third instars and semipupa; drawings of mandibles and maxillae of first, second and third instars. Life cycle in incipient colony: egg 9-12 days, larva 8-14 days, semipupa 5 days, pupa 12-18

1973]

Wheeler & Wheeler — Ant Larvae

77

Fig. 3. 0 cymyrmex arnoldi. a, body hair, X167; b, larva in side view,

Xl6; c, head in anterior view, X81; d, right antenna in anterior view, X376; e, left mandible in anterior view, X133. Fig. 4. Procry ptocerus adlerzi. a, head in anterior view, X67; b-d, three types of body hairs, X267; e, left mandible in anterior view, X185; f, larva in side view, X 14.

7§

Psyche

[March-June

days, total 36-45 days. Life cycle in mature colonies: egg 8-12 days, first instar 2-7 days, second instar 3-7 days, third instar 10-19 days, semipupa 5 days, pupa 12-18 days, total 43-63 days.

Genus Triglyphothrxx Forel

Triglyphothrix striatidens Emery

immature larva. Length (through spiracles) about 2.2 mm. Dorsal profile C-shaped ; ventral feebly sigmoid ; thorax stout and curved ventrally but not differentiated from abdomen in diameter; abdomen bag-like. Spiracles small; diameter diminishing posteriorly. Integument of venter of anterior somites sparsely spinulose, the spi- nules minute and in short transverse rows. Body hairs very few: 6 on Ti, 2 each on T2-AIV; 0.008-0.033 mm long, longest and with multifid-tip on Ti, becoming shorter and simple posteriorly. Head large; cranium subpyriform. Each antenna with three sensilla, each of which bears a spinule. Head hairs minute (about 0.006 mm long) , simple, six only, near mouth parts. Labrum twice as broad as long, bilobed, lateral borders curved; each lobe with five sensilla on the anterior surface, two contiguous sensilla on the ventral bor- der, and one isolated and three contiguous sensilla on the posterior surface; entire posterior surface moderately spinulose, the spinules minute and in short rows which are arranged in longer subtransverse rows medially, laterally the spinules are coarser and isolated. Each mandible heavily sclerotized, narrowly subtriangular in anterior view; of two portions: lateral thick and terminating in a long

slender apical tooth ; medial blade arising from the anterior surface and bearing two sharp-pointed medial teeth. Each maxilla with the apex conoidal and bearing a few spinules; palp a low rounded knob with five sensilla; galea a short frustum with two apical sensilla. Labium feebly bilobed, with minute spinules in subparallel rows; each palp a low rounded knob with five sensilla ; an isolated sensillum between each palp and the opening of the sericteries, the latter a short transverse slit. Hypopharynx densely spinulose, the spinules minute and in short arcuate rows which are arranged in long sub- parallel transverse rows, base with numerous heavily sclerotized long- itudinal ridges. (Material studied: numerous immature larvae from New South Wales, courtesy of Rev. B. B. Lowery.)

Tribe Cryptocerini3

revision : Posterior surface of labrum usually without spinules.

Hypopharynx usually without spinules.

3In 1949, M. R. Smith concluded that the well established name Crypto- cerus, which had been in use for 146 years, was a synonym of Cephalotes ; he changed Cryptocerini to Cephalotini, Cryptocerus to Paracryptocerus and subgenus Cryptocerus to Harnedia. In four pages could anyone intro- duce more confusion into stable nomenclature? The old names should have been conserved. We refuse to accept any of these changes.

1973]

Wheeler & Wheeler — Ant Larvae

79

Genus Cryptocerus Fabricius Cryptocerus rohweri Wheeler

Creighton and Nutting 1965: 63 — “Worker brood developed

from egg to adult in about three months (egg to larva ± 27 days; larva to pupa ± 33 days; pupa to adult ± 23 days).” Eggs de- veloped in about a month into male larvae, which overwintered.

Cryptocerus ( Cyathomyrmex) pallens Klug

immature larva. Length (through spiracles) about 3.4 mm. Similar to C. minutus (called Paracryptocerus rninutus in 1954: 155) except as follows. Head very large and covering approximately half of the anterior end. Integument of venter of anterior somites and all surfaces of posterior somites with a few minute spinules in short transverse rows. Body hairs: (1) 0.006-0.038 mm long, most nu- merous on the prothorax; (2) about 0.225 mm long. Head hairs very numerous, slightly longer (0.013-0.05 mm long). Anterior surface of labrum with eight hairs and/or sensilla. Mandibles moderately to feebly sclerotized. An isolated sensillum between each palp and the opening of the sericteries.

very young larva. Length (through spiracles) about 1.6 mm. Body subellipsoidal, head on the anterior end (i.e., similar to 1.9 mm larva of C. minutus, 1954: 155). Otherwise similar to mature larva except in the following details. Entire integument spinulose, the spinules minute and in short transverse rows. Head hairs shorter (0.013-0.038 mm long). Labrum with ten hairs on the anterior surface. Mandibles sickle-shaped, no blade.

Material studied: five larvae from Brazil, courtesy of Dr. K.

Lenko.

Genus Procryptocerus Emery

revision. Profile cataulaciform. Body hairs moderately numer- ous. Of three types: (1) simple with flexuous tip; (2) tip short- branched, multifid; (3) anchor-tipped. Head hairs of two types: (1) simple, with long flexuous tip; (2) with short-branched (multifid) tip. Mandibles cryptoceriform. Posterior surface of labrum with or without spinules. Maxillae adnate and rounded. Hypopharynx with or without spinules.

Procryptocerus adlerzi (Mayr)

(Fig. 4)

Length (through spiracles) about 4.6 mm. Profile cataulaciform (i.e., straight, elongate-subelliptical ; segmentation indistinct; head applied to the ventral surface near the anterior end) ; no neck. Anus ventral, with a posterior lip. Leg, wing and gonopod vestiges present.

8o

Psyche

[March-J une

Six feebly differentiated somites. Spiracles small ; decreasing in diam- eter posteriorly. Integument of ventral surface of anterior somites with minute spinules in short transverse rows, entire surface of poste- rior somites with scattered minute spinules. Body hairs moderately nu- merous. Of three types: (i) 0.013-0.036 mm long, simple, tip very fine and flexuous; (2) 0.022-0. 13 mm long, tip denticulate, a few on each somite; (3) about 0.2 mm long, with flexuous shaft and anchor- tip, four on each AI-AIV. Cranium subcircular in anterior view. Antennae moderately large; just below middle of cranium; each with three small sensilla, bearing a spinule each. Head hairs moderately numerous, short to moderately long. Of two types: (1) 0.01 9-0.03 mm long, simple, with long fine flexuous tip, the more numerous type; (2) 0.028-0.075 mm long, with short-branched tip. Labrum arcuate; with eight hairs 0.025-0.038 mm long; ventral border with four isolated and two clusters of three sensilla each ; posterior sur- face with four isolated and two clusters of three sensilla each; spi- nules lacking. Mandibles cryptoceriform (i.e., stout, subtriangular in anterior view; lateral portion thick and terminating in a sharp-pointed apical tooth ; medial blade arising from the anterior surface and bear- ing two subapical teeth, which are subequal to apical tooth). Max- illae rounded and adnate; each palp a short peg with five sensilla, larger than the galea, the latter a short cylinder with two apical sensilla. Labium small ; each palp a short stout peg with five sensilla ; an isolated sensillum between each palp and the opening of the seric- teries, the latter a short transverse slit. No spinules on hypopharynx.

immature larva. Length (through spiracles) about 3.1 mm. Body relatively stouter. Head on anterior end. Integument of dorsal surface of posterior somites with rather coarse spinules. Body hairs (1) 0.025-0.05 mm long; (2) 0.0 19-0. 125 mm long; (3) about O.225 mm long, four on the dorsal surface of each AI-AIV. Other- wise similar to mature larva.

very young larva. Estimated length about 1.4 mm. Similar to the immature larva except as follows. Body hairs sparse: (1) 0.002-0.025 mm long; (2) 0.024-0.05 mm long, with tip denticulate to flattened; (3) about 0.086 mm long, on AI-AVI. Head hairs sparse, all short spikes (about 0.003 mm long). Anterior surface of labrum with six hairs about 0.006 mm long. Each mandible with the apex turned medially; all teeth narrowly sharp-pointed. Each labial palp represented by a cluster of five sensilla.

Material studied: 14 larvae from Brazil, courtesy of Dr. K.

Lenko.

1973]

Wheeler & Wheeler — Ant Larvae

81

Procryptocerus regular if Emery

Length (through spiracles) about 5.4 mm. Similar to P. adlerzi except as follows. Body stouter; head relatively larger. Integument with few spinules on the anterior somites. Body hairs (1) 0.013- 0.05 mm long; (2) 0.038-0.1 13 mm long; (3) four on the dorsa of each AI-AIV. Cranium transversely subelliptical. Antennae larger. Head hairs numerous; (1) 0.007-0.05 mm long; (2) 0.038- 0.087 mm long. Labrum feebly bilobed, with eight hairs 0.01 3-0.05 mm long; each lobe with six isolated and two contiguous sensilla on and near the ventral border; entire posterior surface with a few transverse rows of minute spinules. Each mandible with apical tooth longer and basal teeth shorter. Each galea represented by two con- tiguous sensilla. Each labial palp represented by a cluster of five sensilla. Hypopharynx with a few transverse rows of minute spi- nules.

young larva. Length (through spiracles) about 1.8 mm. Sim- ilar to mature larva except as follows. Entire integument with min- ute spinules, in short rows anteroventrally, elsewhere isolated, coarsest posteriorly. Body hairs sparse; (1) 0.003-0.044 mm long; (2) 0.013-0.088 mm long; (3) 0.125-0.188 mm long. Head hairs moderately numerous, 0.003-0.044 mm long, simple. Hairs on labrum about 0.003 mm long.

youngest larva. Length (through spiracles) about 1 mm long. Similar to young larva except as follows. Body egg-shaped. Spiracles very small. Integument with spinules on the dorsal surface of abdo- men from spiracle to spiracle, more extensive on AIX and AX. Body hairs very sparse; (1) 0.003-0.025 mm long, few, none on AIX and AX: (2) 0.025-0.075 mm long, few, with short-bifid to short-multifid tip, on lateral surfaces; (3) 0.09-0.18 mm long, four on the dorsum of each AI-AVII, with straight shaft and smooth anchor-tip. Head hairs very few, about 0.003 mm long. Posterior surface of labrum lacking spinules. Each mandible with one apical and one medial tooth. Each maxillary palp represented by a cluster of sensilla. Labium and hypopharynx without spinules.

MALE LARVA. Length (through spiracles) about 5.8 mm. Sim- ilar to worker larva except as follows. Thirteen feebly differentiated somites. Body hairs (3) about 0.15 m m long, four on the dorsal surface of each AI-AIV. Cranium subrectangular. Head hairs (1) 0.013 mm long; (2) about 0.38 mm long, with short-bifid tip, about eight on the cranium. Mandibles heavily sclerotized.

Material studied : numerous larvae from Brazil, courtesy of Dr.

K. Lenko.

82

Psyche

[March-J une

Procryptocerus striata scabriuscula Emery Length (through spiracles) about 5.2 mm. Similar to P. adlerzi except as follows. Body stouter. Integument of venter of anterior somites and dorsa of posterior somites with minute spinules in short transverse rows. Body hairs (1) 0.003-0. 19 mm long, spike-like; (2) 0.013-0.05 mm long, very few on each somite. Cranium trans- versely subelliptical. Head hairs few. Lab rum with six hairs and six isolated and two clusters of three sensilla each on the anterior surface; posterior surface with six isolated and two clusters of two or three sensilla each and with minute spinules in short arcuate rows, the rows forming a reticulate pattern. Mandibles quadrilateral, heavily sclerotized, with all teeth straight and round-pointed. Each maxillary palp an ungula, with two apical and three lateral sensilla; each galea a very short stout peg with two apical sensilla. (Material studied: nine larvae from Mexico, courtesy of Roy R. Snelling.)

Literature Cited Bruder, K. W., and A. P. Gupta

1972. Biology of the pavement ant, T etramorium caespitum . Ann. Entomol. Soc. Amer. 68: 358-367.

Creighton, W. S.

1965. The habits and distribution of Macromischa subditiva Wheeler. Psyche 72: 282-286.

Creighton, W. S., and W. L. Nutting

1965. The habits and distribution of Cryptocerus roh'weri Wheeler. Psyche 72: 59-64.

Kempf, W. W.

1959. A synopsis of the New World species belonging to the Nesomyr- mex-group of the ant genus Leptothorax Mayr. Studia Entomol. (Rio de Janeiro) 2: 291-432.

Smith, M. R.

1949. On the status of Cryptocerus Latreille and Cephalotes Latreille. Psyche 56: 18-21.

1950. On the status of Leptothorax Mayr and some of its subgenera. Psyche 57: 29-30.

Urbani, C. B.

1968. Studi sulla mirmecofauna d’ltalia. IV. La fauna mirmecologico delle isole maltesi ed il suo significato ecologico e biogeografico. Ann. Mus. Civ. Stor. Nat. Genova 77: 408-559.

Wheeler, G. C., and Jeanette Wheeler

1954. The ant larvae of the myrmicine tribes Cataulacini and Cepha- lotini. J. Washington Acad. Sci. 44: 149-157.

1955. The ant larvae of the myrmicine tribe Leptothoracini. Ann. Entomol. Soc. Amer. 48: 17-29.

1960a. Supplementary studies on the larvae of the Myrmicinae. Proc. Entomol. Soc. Washington 62: 1-32.

1960b. The ant larvae of the subfamily Myrmicinae. Ann. Entomol. Soc. Amer. 53: 98-110.

A NEW SPECIES OF ANACIS FROM NORTHWEST ARGENTINA (HYMENOPTERA, ICHNEUMONIDAE)

By Charles C. Porter*

Department of Biological Sciences, Fordham University,

Bronx, New York 10458

In two previous contributions (Porter 1967, 1970), the author characterized the mesostenine ichneumonid genus Anacis, assigning to it four species from Chile and contiguous regions of southwestern Argentina. Meanwhile, Townes (1969, p. 176-177), as a result of his study of the world Mesostenini, enlarged the definition of A nacis to include also Cryptus exul (Turner, 1919, Ann. & Mag. Nat. Hist. (9)3: 558) from Tasmania. Consequently, Anacis seemed to emerge as pertaining to that zoogeographic category comprised of taxa restricted at the present time to the Nothofagus zone of south- ern South America and to similar areas of the Australian region.

Now, however, discovery of a fifth Neotropic Anacis from sub- tropical wet forest in northwestern Argentina obliges us to modify our distributional concept of the genus. Thus, in South America Anacis appears to be of Andean rather than of strictly Neantarctic or Araucanian range and, quite possibly, extends to other areas on the continent. Its New World distribution, therefore, may be com- pared to that of several other ichneumonid genera — such as Macro - grotea , Trachysphyrus (sensu Townes), Picrocryptvides, Dotocryp- tus, Deleboea } Alophophion , and Thymebatis — all of which are well represented in Andean and temperate South America, including Chile, but which concurrently have a greater or lesser number of species on the peripheries of the lowland tropics. Taxa of this same distributional type which moreover have species in the Australian region are, of course, much rarer, but the ichneumonid genus Labena (two species also reach North America) and the seolioid family Thynnidae constitute approximate parallels.

The present study offers a description of this new Argentine Anacis and a revised key to all known South American species of the genus.

KEY TO THE SOUTH AMERICAN SPECIES OF ANACIS

(Based on females)

1. Mesoscutum mat, finely granular; setae of second gastric ter- gite dense, mostly approaching or exceeding the length of their

83

84 Psyche [March-June

interspaces; mesosoma pale red or reddish brown with white markings and black areas of variable extent 2

Mesoscutum silky shining with more or less fine punctuation; setae of second gastric tergite sparser, mostly shorter to much shorter than the length of their interspaces; mesosoma black with sparse to profuse white markings 3

2. Gaster black with white apical bands on tergites; flagellum with

a white annulus; hind-tarsomeres 2-4 white; first flagellomere 8.5-10.0 as long as deep at apex; postpetiole weakly expanded, O.8-O.9 as wide apically as long from spiracle to apex; ovi- positor tip 0.18-0.20 as high at notch as long from notch to

apex 1. A. f estiva Porter

Gaster bright reddish brown with more or less prominent white apical bands on tergites; flagellum without a white annulus; hind-tarsus without white markings; first flagellomere 6. 2-7. 3 as long as deep at apex; postpetiole more strongly expanded, 1. 3-1. 4 as wide apically as long from spiracle to apex; ovipos- itor tip 0.26-0.31 as high at notch as long from notch to apex 2. A. tucumana n. sp.

3. Thorax and propodeum with profuse white markings; all gastric

tergites with a complete white apical band ; apical margin of clypeus with a small subdentate median projection; dorsal mar- gin of pronotum without a definite submarginal groove

3. A. stangeorum Porter

Mesosoma with white at most on anterior margin of pronotum, tegula, and subalarum; not all gastric tergites with a com- plete white apical band; no median projection on apical margin of clypeus; dorsal margin of pronotum with a conspicuous submarginal groove 4

4. Legs mostly black with white markings; gaster with the follow-

ing white: median subapical mark on tergite 1, sometimes

marks on 2 and 3, and broad apical bands on 4-7 ; sheathed portion of ovipositor 0.5-0.6 as long as fore-wing; nodus of

ovipositor tip with an unusually large and deep notch

4. A. varipes Porter

Legs mostly orange; gaster with white at most on tergites 6-8 and only on 7 sometimes with a complete white apical band; sheathed portion of ovipositor O.3-0.4 as long as fore-wing; nodus with a small but distinct notch 5. A. rubripes Spinola

1973]

Porter — A nacis

85

2. Anacis tucumana new species

( fig. 0

Holotype: female, ARGENTINA ( Tucuman : Horco Molle,

Dto. Tafi, October 24, 1970, C. C. Porter). (Tucuman). Para- types: 4 females, ARGENTINA ( Jujuy : Post a de Lozano, Octo- ber 26, 1969, December 8, 1969, C. C. Porter; Tucuman: Horco Molle, Dto. Tafi, September 9-1 1 & September 30, 1969, C. C. Porter). (Gainesville, Porter, Tucuman) .

Female: Color: flagellum pale reddish brown with slight dusky

staining, especially above on first segment; pedicel largely blackish brown above and pale reddish brown to yellowish below; scape largely blackish brown above and grading through reddish brown into yellowish below; head black with some brown staining toward apex on mandibles, sometimes in malar space, often around clypeus above, often irregularly on face, and on antennal sockets, as well as with the following white: most of basal 1/2 of mandible; very large transverse blotch on clypeus; and a complete or sometimes ventrally interrupted orbital ring, which is broadest below where it extends 1/2 to 3/4 or more the distance into malar space; mesosoma bright reddish brown with black on areas of variable extent includ- ing most of prothorax (which at most is irregularly reddish stained), mesoscutum slightly to entirely, mesopleuron slightly to on as much as dorsal 1/3 plus most of prepectus, many margins and sutures more or less broadly, as well as with the following white: broad anterior margin of pronotum except toward apex ; broad dorsal mar- gin of pronotum except for about median 1/4-1/3; tegula; most of subalarum; sometimes large spot toward lower anterior corner of mesopleuron; small area in upper hind corner of mesopleuron on mesepimeron; most of apical 3/4-778 of scutellum; and sometimes most of postscutellum ; gaster bright reddish brown with a prominent to dull and little contrasting white apical band on tergites 1 or 2- 8 ; fore-leg with coxa mostly white with a conspicuous dorso-apical brown blotch and some brown staining postero-basally ; trochanter white with a large pale brown blotch dorsally; trochantellus white with dusky to black staining on apex and above; femur bright red- dish brown grading into whitish below and with a little dusky to black staining on base; and tibia and tarsus somewhat duller brown with some faint dusky staining and with fifth tarsomere mostly dusky to blackish; mid-leg with coxa pale reddish brown with a large, pale to dark brown dorso-apical blotch and sometimes a less well defined ventral brownish area basad, as well as with white on

86

Psyche

[March-June

Fig. 1. Anacis tucumana Porter, female holotype. Lateral view of mesosoma and gaster, showing color pattern.

most of dorso-basal 3/5 and on an extensive ventro-apical area; trochanter white with a large brown area dorsally; trochantellus white basally and below and blackish brown apically and above; femur bright reddish brown with base narrowly blackish; and tibia and tarsus a little duller reddish brown with last tarsomere dusky to blackish; hind-leg with coxa uniformly bright red brown, except sometimes for an obscure whitish area above near base; trochanter varying from white with reddish brown staining to almost uniformly reddish brown with white narrowly on apex, and sometimes with blackish l rown staining on about basal 3/4 above; trochantellus more or less blackish brown above and mostly reddish to brownish white or white below; femur bright red-brown with base narrowly blackish; and tibia and tarsus a little more dully red-brown with last tarsomere dusky to blackish; wings hyaline with very slight dusky staining apicad. Length of fore-wing: 5. 3-5. 7 mm. 1st fiagel- lomere: 6. 2-7. 3 as long as deep at apex. Clypeus: in profile high and asymmetrically convex to bluntly subpyramidal, with the apical face shorter than the basal and a little concavely declivous ; the apical margin practically straight, not produced or dentate medially. Malar space : 0.75-0.85 as long as basal width of mandible. Temple: 0.30- 0.36 as long as eye in dorsal view; gently rounded-off and strongly

1973]

Porter — A Jiacis

87

receding. Fore-tibia: moderately stout but scarcely swollen. Pro- notum: dorsal margin a little swollen, especially anteriad, and with- out a submarginal groove; epomia sharp in scrobe but only faintly prolonged below; anterior margin not angled at mid-height below. Mesoscutum: notauli traceable about 1/2-2/3 the length of meso- scutum, well defined but not very sharp anteriad and becoming much weaker behind ; surface mat and finely granular with abundant, small, faint, subadjacent to confluent punctures and an area of con- trastingly coarser puncto- reticulation apicad between and beyond notauli. Mesopleuron: subalarum not unusually swollen or expanded; speculum swollen, mostly smooth and polished; surface otherwise more or less strongly shining with fine but moderately strong, rather uniform, obliquely longitudinal wrinkling, which becomes only slight- ly more irregular on lower 1/2, and with medium-sized, variably distinct intercalated punctures which are best defined on lower 1/2. Wing venation: radial cell 2. 8-3. 2 as long as broad; areolet large and broad, intercubiti strongly convergent above, 2nd abscissa of radius 0.6-0.7 as long as 1st intercubitus; 2nd recurrent vertical, at most weakly outcurved on upper 1/2; disco-cubitus broadly angled or simply arched, without or rarely with a vestige of a ramellus; nervulus at least slightly antefurcal ; mediella strongly arched ; axil- lus close to hind-margin of wing. Propodeum: moderately short and high in profile, basal face gently arched and sloping rearward to join the much more steeply declivous but not sharply discrete, subequal apical face ; spiracle round ; basal trans-carina sharp throughout, weakly to moderately bowed forward medially, rather far from base of propodeum; apical trans-carina more or less traceable throughout, but becoming a little to, often, very weak and irregular on its broad- ly bowed forward median portion, laterally well defined, forming very low, sub-crescentic cristae and continuing ventrad to pleural Carina; areola not defined; without lateral longitudinal carinae; surface basad of basal trans-carina shining with considerable fine wrinkling, especially mesad, and abundant, rather large, shallow, subadjacent or sparser to confluent punctures but distad of basal trans-carina uniformly mat with stronger, granularly reticulate wrinkling and puncto-reticulation. 1st gastric segment: with a low and weakly crescentic lateral flange at base; petiole moderately broad and flat; postpetiole strongly expanded and 1.3-1.4 as wide apically as long from spiracle to apex; ventro-lateral carina sharp on postpetiole and about apical 1/2 of petiole but sometimes more or less fading out toward base on petiole, or continuing sharp to base; dorso-lateral carina fine and sharp for a short distance near spiracle

88

Psyche

[March-June

and again toward base of petiole but otherwise faint or absent; dor- sal carinae absent or at most faintly suggested above spiracle. Cas- ter: moderately elongate fusiform; 2nd tergite dully shining to mat with fine, granularly reticulate wrinkling and abundant, medium sized to large, mostly obscure, densely intercalated punctures which emit numerous short setae that in great part approach or equal the length of their interspaces; the following tergites with somewhat longer and denser setae that mostly equal or exceed the length of their interspaces. Ovipositor: sheathed portion 0.2 as long as fore- wing; straight, moderately stout, strongly compressed; nodus high, with a small but sharp notch ; dorsal valve in profile with a straight to slightly concave taper between notch and apex; ventral valve on tip with fine, inclivously oblique ridges; tip 0.26-0.3 1 as high at notch as long from notch to apex.

Male*, unknown.

Types: The holotypes and two paratypes are deposited in the

collection of the Institute Miguel Lillo, San Miguel de Tucuman, Republica Argentina. One paratype has been donated to the Florida State Arthropod Collection (Gainesville, Florida, USA) and a fourth paratype is in the collection of Charles C. Porter (RFD 3, Cambridge, Maryland, USA).

Discussion: Among South American species of its genus, Attach

tucumana comes closest to the Araucanian A. f estiva, as shown by several common characters emphasized in the key. Nonetheless, that relationship is comparatively remote and tucumana has some features which set it apart from the other South American Anacis ; for ex- ample, its shorter notauli, shorter second radial abscissa, less elongate propodeum, weaker dorso-lateral and dorsal carinae of the first gas- tric tergite, and slightly shorter ovipositor. Indeed, the northwest Argentine and Araucanian populations of Anacis probably have been out of contact since the late Tertiary and early Pleistocene Andean uplift, although it may be surmised that this is an old genus which ranged throughout the climatically and biotically more uniform South America of Pre-Andean times.

Worth noting also is the difference in abundance between the southern and northern Anacis. At least two of the southern species (A. f estiva and A. ruhripes ) are very common insects likely to be encountered in numbers almost any day during the growing season; whereas, A. tucumana only has been collected on five occasions. This circumstance coincides with the probable relict status of Anacis. Populations isolated in the Araucanian zone, where almost none of

1973]

Porter — A nacis

89

the modern Neotropic ichneumonid fauna has penetrated, would be expected to flourish in the absence of aggressive competitors; whereas, populations in an area such as the Selva Tucumano-Boliviana, where occur scores of the Neotropic genera represented by hundreds of species, would form a much more inconspicuous part of the fauna and might have difficulty surviving at all.

Field notes: All localities for this species belong to the wet

subtropical forest community commonly designated Selva T ucumano- Boliviana. Horco Molle in Tucuman Province is located in the lower stratum of this forest (about 700 m.), while Posta, de Lozano in Jujuy Province is at 1600 m. in an area of transition between several types of environment, including many components of the Selva and some Chaco elements, as well as stands of alders ( Alnus joruliensis) .

Specimens of A. tucumana were collected by sweeping in weedy areas both at the edge of the forest and in partially shaded places within the forest.

Acknowledgements

Part of the material covered in this study was collected while the author was working as Associate Investigator under a United States National Science Foundation Grant (GB-6925) awarded to Dr. Howard E. Evans of the Museum of Comparative Zoology at Har= vard University.

The figure was inked by Miss Alicia Sandoval of the Instituto Miguel Lillo.

References

Porter, C. C.

1967. A Review of the Chilean genera of the tribe Mesostenini. Studia Ent. 10: 369-418.

1970. The Genus Anacis in Argentina. Acta Zoologica Lilloana 26(2) : 9-22.

Townes, H. K.

1969. Genera of Ichneumonidae, Part 2, Gelinae. Mem. Amer. Ent. Inst. 12.

GROWTH OF THE ORB WEAVER, ARANEUS DIADEM A TVS, AND CORRELATION WITH WEB MEASUREMENTS*

By Jay Benforado and Kent H. Kistler Division of Research, North Carolina Department of Mental Health Raleigh, North Carolina 27611

Introduction

It is a well-known fact that within any population of spiders of similar age there is considerable variation in the size of individual spiders of the same species. In literature as early as 1890, McCook has observed this variation and repeated observations (Comstock, 1940; Savory, 1928) have verified this phenomenon. Although ob- servations are frequent, explanations are few. Bristowe (1958) cites differences in feeding as a reason for differential size, but the reference is made merely in passing and to the authors’ knowledge is not elaborated upon elsewhere. This paucity of explanation lends itself to further analysis of the factors contributing to the phenom- enon of differential size.

Our purpose in this paper is to isolate some of the factors which contribute to differential size in Araneus diadematus Clerck (for identification of species, see Levi, 1971), and to elaborate upon cer- tain of these factors as we are able.

Corresponding with differential size, in an orb-weaver such as Araneus diadematus , differential growth is also manifested in chang- ing dimensions of the web. That large differences in dimensions exist betwen the individual webs of spiders is also a well-known fact. An attempt to clarify some of the factors influencing web changes is also made.

Method

environment: The spiders used were from two cocoons of Ara-

neus diadematus , obtained from Canastota, New York, which hatched on April 26, 1972. From the time of hatching and throughout the experiment, the spiders were kept in a room which was lighted 16 hours per day aand kept cool during the eight dark hours by an air conditioner. (See Witt, 1971).

early rearing: At the time of hatching, the offspring from each

* Manuscript received by the editor March 1 , 1973

90

1973] Benforado & Kistler — Araneus diadematus 91

cocoon were placed in a separate rearing box. The spiderlings were kept in these boxes, living on a communal web with a constant sup- ply of loose gnats in the box, until they began to build individual webs approximately three weeks after hatching. As each animal built her first web she was removed from the rearing box and placed in an individual glass tube, approximately 1X7 cm, with the ends of the tube stoppered with cotton. From the time the animals were placed in the tubes until onset of the experiment they were fed ap- proximately 10-15 gnats per week, by placing the gnats in the tube with the spiderling. The animals were watered by wetting the cot- ton with water daily.

distribution: Seven weeks after hatching the two sets of spider-

lings were each separated into three equal groups by means of a random numbers chart. No attempt was made to distribute males and females evenly. Although the growth (body weight) of males and females differs, it has been shown that the early growth of both sexes is alike (Witt et al., 1972). Because of the short duration of the experiment and the difficulty in identifying male spiders before the last molt, distribution of males and females was left to chance.

At the time of initial grouping the two sets of spiderlings num- bered twenty and thirty respectively. It was decided to feed each of the three groups of each set according to a different feeding sched- ule: one group every day, one group twice weekly, and one group every ten days. Thus there were six groups, one for each set of offspring on each feeding schedule. After one week of this procedure, however, it was decided because of the small size of the groups to reduce the number of schedules to two, and the middle /schedule was dropped and its members distributed randomly between the lighter and heavier-fed groups.

Data for animals that died or escaped during the course of the experiment were removed, so figures represent only animals observed for the duration of the experiment.

weighing: Each spider in the heavy-fed groups was weighed once

a week, to 0.1 mg, while animals in the light-fed groups were weighed on the day of feeding and the day after feeding.

web analysis: After eleven days of controlled differential feeding

in the tubes, the spiders were transferred to aluminum and glass laboratory cages, 50 X 50 X 10 cm. At this time the animals were eight weeks old. From this time on the spiders began to build webs. Photographs of webs were taken daily and analyzed (Reed et al., 1965). Daily records of web building were kept and the webs were

92

Psyche

[March-June

Figure 1. Mean weights of 19 light-fed and 15 heavy-fed Araneus diadematus with standard errors (vertical lines). Figures are for both sets of spiderlings combined. Sharp increases in weight in the light-fed group are due to the animals being weighed before and after feeding. Note the increasing difference in weight between the light-fed and heavy- fed group.

1973]

Benforado & Kistler — Araneus diadematus

93

destroyed daily with the thread left in the cage for the spider to digest.

feeding: While in the tubes, the spiders were fed by placing a

previously weighed de-winged housefly in the tube daily or every ten days. Those spiders that would not eat a housefly had three to seven unweighed gnats placed in their tube. By visual inspection the following day it was determined whether the fly had been eaten. The remains of the eaten flies were then weighed to obtain an approxi- mation of the amount eaten by each spider. The spiders were watered by wetting the cotton every other day.

After being placed in the cages, if the spider had a web, feeding was by means of placing the housefly in the web ; if there was no web, we attempted to induce the animal to eat by placing the fly in front of its mouth. The heavy-fed spiders were offered at least one fly per day and more, if they would accept it. The light-fed group was fed one fly once every ten days. If on the day of feeding of the light-fed group a spider would not eat, a note was made and the attempt repeated until successful. All spiders were watered on Mondays, Wednesdays and Fridays by spraying a small amount of water in each cage.

molts: From the onset of the experiment molts were recorded by

date of the molt to give an indication of the maturation of the animal.

Results

feeding AND weight increase: At the end of a period of five

weeks the two feeding schedules resulted in two significantly differ- ent weight groups. This development is shown in Figure i, which illustrates the increasing difference in weight between the two groups. At the onset of differential feeding the mean weights of the two groups were alike, however, a T-test between the mean weights at the end of the experiment is significant at the o.i % level.

An analysis of covariance was performed on the data.. Because the original data was non-homogeneous, a transformation [log (x + io) ] was made (Winer, 1962). The initial observation was used as a covariate in the analysis of covariance. Because the analysis of covariance indicated no significant difference in the behavior (growth) of the two families, all figures are for both families com- bined. For the heavy-fed group the mean weight changed from 7.93 mg db 1.04 on June 12 to 74.28 mg ± 10.93 on July 17. The mean weight of the light-fed group changed from 6.40 mg dz 0.98

94

Psyche

[March-June

on June 12 to 17.91 mg ± 2.56 on July 13; there was a significant interaction between time and feeding schedule below the 1% level. FEEDING AND Maturation: If the number of molts over time is

taken as an indication of speed of maturation, then a relationship between feeding and rate of maturation can be seen. During the period of differential feeding the number of molts of the heavy-fed and light-fed groups differed significantly at the 5% level. The heavy-fed group had a mean number of 3.0 molts while the light-fed group had a mean number of 2.3 molts. These results are in agree- ment with the findings reported by Deevey (1949) with Latrodec- tus mactcins (Fabricius) and indicate that in the laboratory with only food quantity as a variable, a relationship exists between the rate of weight increase and the rate of maturation.

INITIAL WEIGHT and rate of growth: From the beginning of

the experiment we noted a wide variation of weights of the individ- ual animals. At the onset of differencial feeding individual weights ranged from 1.1 mg to 16.2 mg. In both the light-fed and heavy- fed groups there existed a positive correlation between initial weight and final weight. For the light-fed group r = 0.7713 and for the

TABLE 1

Measurement	Light-fed Early Late	Heavy-fed Early Late
Mean wt. (mg) of spiders	12.52	22.30	20. 1 4	59-34
Spiral area (cm2)	1 18.92	119.51	1 19.88	138.32
Center area (mm2)	7 1 1 .00	877.53	920.30	1424.30
Thread length (m)	7-35	7-47	7.56	8.47
Mesh width (mm2)	20.16	22.34	21.79	27.48
Angle regularity	4-25	4.16	5.52	4.62
# of oversized angles Relative deviation of	1.67	1.87	2.50	1.80
spiral turns (South)	0.34	0-33	0.41	0.35

Selected measurements of webs built by a group of light-fed and heavy-fed spiders. Because not all spiders built on the same day, early and late webs of both groups were chosen from two five day periods two weeks apart. Measurements are divided into those which measure size (above the broken line) and those which measure regularity. Note the difference between the light-fed and heavy-fed animals in measures of web size at the late date. While the heavy-fed group increased in all size measures (see Fig. 2), no web regularity measures changed. For an explanation of web measure- ments see Witt et al., A Spider’s Web.

1973]

Benforado & Kistler — Araneus diadematus

95

Figure 2. Selected web samples from two spiders: one heavily-fed spi- der and one light-fed spider. Webs are from the periods measured in Table 1 and are all reproduced to the same scale. Note that while both the heavy-fed and light-fed animals began with webs of similar size, after two weeks of differential feeding the large, heavy-fed spiders’ webs had increased in size while the webs of the small, light-fed spiders remained the same size.

96

Psyche

[March-June

heavy-fed group r = 0.9m; both correlations are significant at the 0.1% level. In most instances those animals with the extreme weights at the beginning of feeding remained the extremes in their group. Reasons for the variation in initial weight are unknown. Different rates of growth for light and heavy hatchlings have re- cently been shown to occur in several species of spiders, apparently independent of food available, and appear correlated with different lengths of life (Reed and Witt, 1972).

AMOUNT EATEN : An approximation of the amount eaten was ob-

tained for a three week period. For nine heavy-fed animals the mean amount eaten during the three week period was 115.8 mg and for fifteen light-fed animals the mean amount for the same period was 35.O mg. Within each group, however, there was an enormous variation in the amount consumed : in the heavy-fed group the

amount eaten by individual animals ranged from 209.0 mg to 42.6 mg while in the light-fed group the amount eaten ranged from 49.6 mg to 3 gnats weighing 18 mg.

feeding and web changes: Table i gives a summary of web

changes that accompanied the growth of the animals. In measure- ments of web size both groups increased, with the heavy-fed group having a much larger increase as illustrated in Figure 2. In measure- ments of web regularity both heavy-fed and light-fed groups re- mained constant, as shown in figures of Table 1.

Discussion

The observed differential growth and development in Araneus diadematus seems to be a function of several factors. Although an exposure to a greater than normal supply of food generally results in faster than normal growth and development, even within a group exposed to the same food supply there seems to be a great variation in growth rates. Evidence of these differences is expressed in the increasing standard errors in Figure 1, and seems to be dependent upon individual factors in the animals rather than environmental variations. Large differences in the amount eaten by individual animals in the laboratory existed and presumably exist in nature. These differences seem to correspond to differences in the rate of growth in agreement with the findings of Turnbull in other species of spiders (Turnbull, i960, 1965). However, whether these dif- ferences in the amount of food eaten are due to differences in pro- ficiency in prey-catching or to differences in appetite or some other factor in the animal is not clarified by our findings.

1973]

Benforado & Kistler — - Araneus diadematus

97

Another important factor influencing differential growth is the initial weight of the animal. Variations in initial weights within a family are generally retained during the course of development. Al- though several possible reasons for different initial weights within a family have been given by others, the authors are reluctant to offer any explanations.

In an orb web weaving spider such as Araneus diadematus the amount of food available to the animal is roughly equivalent to the number of prey which become entrapped in the web. The number of prey entrapped in the web is in turn determined by a number of variables such as web-site, size and fine structure of the web, and frequency of web building. Thus, it can be seen that the interaction of the variables resulting in differential size and growth is complex and can be divided into those factors which influence the amount of food available to the spider and those factors which influence the spider’s use of the food available to it.

Repeated attempts have been made to explain web characteristics in terms of characteristics of the individual spider (Peters, 1936). More common, however, has been the notation of changes in the form of the orb web during the life of the spider (Tilquin, 1942; Savory, 1952) and the attempt to relate these changes to changes in the animal (Witt and Baum, i960; Witt, 1963; Reed et al.> 1969). Because influencing factors vary concurrently, it is frequently diffi- cult to assess the causes of changes in the form of the web.

In our experiment we attempted to isolate the effect of one vari- able (weight) while minimizing the effect of a variable which nor- mally changes concurrently (time). All animals used hatched on the same date, however, one group (the heavy-fed) gained consid- erable weight over the period measured. The web changes accom- panying these weight increases are summarized in Table 1. Be- cause all of the animals were hatched on the same date, we conclude that increases in web size are due to differences in size of the ani- mals resulting from differential feeding rather than differences in age. If appetite were a factor influencing web size, it would appear that the hungrier, light-fed animals would build a larger web in an attempt to catch more food ; however, this is not the case.

The relationship between food and the web of a spider is a deli- cate one. Without food, the spider’s web-building ability diminishes, but without a web there is no food (Peakall, 1968). Thus, like a businessman, the spider faces the law of diminishing returns. It appears that the hungry spider chooses to conserve its resources

98

Psyche

[March-June

rather than gamble on a larger web trapping more food. Early food deprivation experiments (Witt, 1963) show that the spider con- tinues to build the same size web when deprived of food, but with less thread until finally a decreasing in web size occurs. Because our hungry (light-fed) animals were kept on a diet closer to a main- tenance level than a deprivation level, we observed no decreases in web dimensions.

Feeding conditions in a natural environment vary more than those imposed in a controlled laboratory. Yet the spider is able to survive in these naturally diverse conditions because of its adaptability. In situations where there is little food available, the spider is able to survive by growing at a slow rate and maintaining the same size web. Where food is abundant, the spider takes advantage of the situation, growing at a fast rate and increasing the size of its web.

The spider has developed a method for coping with a wide range of feeding conditions. By varying its body and web growth, the spider can survive under the diverse conditions imposed by nature, thus minimizing the necessity of seeking new food supplies and re- locating the web. Our findings provide new insight into the spider as an example of an animal that adapts itself successfully to its en- vironment.

Summary

Spiders from two cocoons of Araneus diadematus were exposed to five weeks of two different feeding schedules: one group was of- fered large amounts (one housefly per day) of food, the other group scarce (one fly every ten days) amounts. Although both groups in- creased in weight, weight gains of the heavy-fed group were signifi- cantly greater than those of the light-fed group, regardless of cocoon origin. Within each group there was a wide variation in the growth of individual animals, indicating the presence of factors other than food supply; i.e. animals with extreme weights within a group at the onset remained the extremes.

In conjunction with increases in weight, over the three week period of observation, webs of the heavy-fed spiders showed an increase in size but not in regularity and shape in comparison to webs of the smaller, light-fed animals of the same age which did not change. Such data suggest an increased chance of survival of the species through variations in rate of growth and maturation dependent on environmental factors.

1973]

Benforado & Kistler — Araneus diadematus

99

Acknowledgements

This work was carried out in the laboratories of the North Caro- lina Department of Mental Health and was supported by Grant Number GB 25274 from the National Science Foundation to Peter N. Witt. The authors gratefully acknowledge the assistance of Dr. Peter N. Witt during all stages and the assistance of Mrs. Mabel B.

Scarboro during the period of laboratory work.

References Cited

Bristowe, W. S.

1958. The World of Spiders. Collins, London.

Comstock, J. H.

1940. The Spider Book. Revised and edited by W. J. Gertsch. Corn- stock, Ithaca, N.Y.

Deevey, G. B.

1949. The development history of Latrodectus mactans (Fabr.) at dif- ferent rates of feeding. Amer. Midland Naturalist. 42: 189-218.

Levi, H. W.

1971. The Diadematus group of the orb-weaver genus Araneus north of Mexico (Araneae: Araneidae). Bull. Mus. Comp. Zool., 141: 131-179.

McCook, H. C.

1890. American Spiders and Their Spinningwork. Vol. 2, Published by the author, Philadelphia.

Peakall, D. B.

1968. The spider’s dilemma. New Scientist, pp. 28-29.

Peters, H. M.

1936. Studien am Netz der Kreuzspinne (Aranea diadema.) 1. Die Grundstruktur des Netzes und Beziehungen zum Bauplan des Spinnenkorpers. Z. Morphol. Okol Tiere, 32: 613-649.

Reed, C. F. and P. N. Witt

1972. Growth rate and longevity in two species of orb-weavers. Bull. Brit. Arach. Soc., 2: 111-112.

Reed, C. F., P. N. Witt and R. L. Jones

1965. The measuring function of the first legs of Araneus diadematus Cl. Behavior, 25 : 98-119.

Reed, C. F., P. N. Witt and M. B. Scarboro

1969. The orb web during the life of Argiope aurantia (Lucas).

Develop. Psychobiology, 2: 120-129.

Savory, T. H.

1928. The Biology of Spiders. Sidgwick and Jackson, London.

1952. The Spider’s Web. Frederick Warne and Co., London and N.Y.

Tilquin, Andre

1942. La Toile Geometrique des Araignees. Presses Universitaires de France, Paris.

Turnbull, A. L.

1960. Quantitative studies of the food of Linyphia triangularis (Clerck) (Araneae: Linyphiidae) . Canad. Ent. 94: 1233-1249.

I <30

Psyche

[March-June

Turnbull, A. L.

1965. Effects of prey abundance on the development of the spider Agelenopsis potteri (Blackwell) (Araneae: Agelenidae). Canad. Ent. 97: 141-147.

Winer, B. J.

1962. Statistical Principles in Experimental Design. McGraw-Hill, N.Y. pp. 606-615.

Witt, P. N.

1963. Environment in relation to behavior of spiders. Arch. Environ. Health, 7 : 4-12.

1971. Instructions for working with web-building spiders in the lab- oratory. BioScience, 21: 23-25.

Witt, P. N. and Ricarda Baum

1960. Changes in orb webs of spiders during growth. Behavior, 16: 309-318.

Witt, P. N., J. O. Rawlings and C. F. Reed

1972. Ontogeny of web-building behavior in two orb-weaving spiders. Am. Zoologist, 12: 445-454.

Witt, P. N., C. F. Reed and D. B. Peakall

1968. A Spider’s Web. Springer Verlag, Berlin.

THE COCKROACH GENUS CALOLAMPRA OF AUSTRALIA WITH DESCRIPTIONS OF NEW SPECIES (BLABERIDAE)

By Louis M. Roth1 and Karl Princis2

Princis (1963) listed 23 species of Calolampra, thirteen of them from Australia and the others from Africa, India, Burma, China, Sarawak and the Philippines. Calolampra simlansis from the Hima- layas was described by Baijal and Kapoor (1966). Recently, the African species were placed in a new genus Pseudocalolampra (Roth and Princis, 1971) and Calolampra laevis (Brunner v. W. ) from Burma and China was assigned to the new genus Calolamprodes (Bey-Bienko, 1969). Of the other species, ‘‘Calolampra” brunneri (Brancsik) is a mislabeled Derocalymma cruralis, and “C.” punctosa (Walker) belongs to the genus Laxta (Princis, 1967, p. 7°8)- Calolampra truncata (Brunner v. W.) was listed by Princis (1967, p. 627) under the genus Rhabdoblatta. In this paper we describe 26 species of Australian Calolampra , of which 12 are new. Five species, C. gracilis (Brunner v. W.), C. atomifera (Walker), C. notabilis (Walker), C. signatura (Walker), and C. propria (Walker) have been considered to be synonyms of C. irrorata (Fabr. ) (Princis, 1963). Of these, propria is a synonym of C. atomifera (Walker) ; the other 4, we believe, are valid species.

The genitalia of Calolampra consist of 3 major phallomeres (Fig. 1). An L2d (dorsal sclerite of the second left phallomere) is apparently absent. The prepuce is covered with microtrichia, curves upward and is fused to the L2vm. When well developed the prepuce is cup-shaped and with L2vm as a handle looks like a ladle; rarely is the prepuce unmodified (e.g. Fig. 82). Li (first sclerite of left phallomere) and R2 (hooked sclerite of the right phallomere) is well developed and lacks a subapical incision. The hypandrium (subgenital plate) is symmetrical and bears 2 small equal sized styli. All photographs of genitalia, supra-anal plate, and hypandrium are from specimens treated with 10% KOH, dehydrated, and mounted in Permount. Supra-anal plates were mounted dorsal sur- face upward and subgenital plates ventral side uppermost. One or

’Pioneering Research Laboratory, U.S. Army Natick Laboratories, Natick, Massachusetts 01760.

Zoological Institute, Lund University, S-223. 62 Lund, Sweden.

Manuscript received by the editor March 26, 1973.

IOI

102

Psyche

[March-June

Figure 1. Calolampra confusa. Male genitalia, dorsal view (115 MCZ), A. (without exact locality). LI — first sclerite of left phallomere; L2vm — ventromedial sclerite of second sclerite of left phallomere; P